EVOLUTION OF LAND PLANTS (UNDER CONSTRUCTION)

Background. As little as thirty years ago our understanding of the evolution of land plants was very different from what it is now. Then it was usually thought that Psilotum represented a very ancient group, often being compared with plants from the Devonian Rhynie Chert; that horsetails were also ancient (but not immediately related), and were linked to the fossil Sphenophyllum and its relatives; and that Lycopodium, Selaginella and relatives formed a group, after which came ferns, gymnosperms, and seed plants. Bryophytes, that is mosses, liverworts and hornworts, represented the earliest land plants. Relationships between the bryophytes was unclear, but they seemed to form a group from which other land plants evolved, while the aquatic Characeae (inc. Nitella) with their quite complex haploid plant bodies were thought to be the algal group most closely related to land plants (e.g. Graham 1993).

Before going any further, note that when describing relationships in the context of a branching phylogenetic tree, the use of terms like "high", "low", "basal", etc., to describe aspects of the branching pattern may seem perfectly appropriate, but at one level, even talking about branching points of this tree makes little sense; there is no trunk, and of any "branch" (one of a pair of sister clades), one cannot be basal to the other. Although on occasion, I do use the term "basal", this is in the context of a ladderized tree and refers to the branch in which there has subsequently been less diversification. Thus Takakia, Equisetum, Amborella, Acorus, Enkianthus, Humbertia, and many other clades can all be called "basal" from this point of view - but note that this has only the topological connotations I have just mentioned, there is certainly no implication that these taxa lack apomorphies or are "primitive" in any way. Indeed, the terms "primitive" and "advanced" are heavily loaded and I have tried not to use them; "plesiomorphic" and its opposite, "derived" and "apomorphic", are the terms to use when talking about individual characters.

Relatives of Land Plants. Given the relationships just described, evolution seemed to have resulted in a fairly straightforward increase in complexity from "lower" to "higher" plants, but this comforting sequence has been severely challenged over the last twenty years or so. Land plants or Embryopsida are members of a clade embedded in a largely aquatic paraphyletic group, the "green algae" (paraphyletic groups will be denoted by scare quotes); the two together make up Viridiplantae. Viridiplantae can be divided into two main clades, Chlorophyta s. str. and Streptophyta, the latter including "Charophyta s.l." within which Embryopsida are embedded. For general information on the evolution of Chlorophyta and Streptophyta, see e.g. Lemieux et al. (2000), Turmel et al. (2002), Sanders et al. (2003), Waters (2003), Delwiche et al. (2004), Lewis and McCourt (2004), Becker et al. (2009), and Leliaert et al. (2012, a great deal of information, 2016). Lewis and McCourt (2004) emphasize that many clades of "green algae" have terrestrial representatives, of which Embryopsida are merely the most prominent. Leliaert et al. (2016) note a number of similarities in the chloroplast genome between their Palmophyllophyceae, perhaps sister to other chlorophyta, and the streptophyte Mesostigma.

Within Chlorophyta s. str. there are a number of algae that are involved in lichen formation (Trebouxia and its relatives) as well as several ecologically very important marine algae. Volvox, Caulerpa, Ulva and Acetabularia are other Chlorophyta.

STREPTOPHYTA

These include land plants (= embryophytes or Embryopsida) and a subset of freshwater green algae like Mesostigma viride (biciliate), Chlorokybus (but see below for these two), Klebsormidium, Spirogyra, Coleochaete and Chara.

Age. Stem streptophytes have been dated to 1200-725 Ma (Yoon et al. 2004: Chaetosphaeridium vs. the rest), Zimmer et al. (2007) give an age of (963-)725(-587) Ma for the divergence of Chlamydomonas from the streptophytes and Morris et al. (2018) HPD ages of 891-629 Ma (Mesostigma vs. the rest).

Evolution: Divergence & Distribution. An ever-growing number of features formerly thought to be restricted to embryophytes are also being found in the more basal streptophytes, if in a somewhat different form/context there (e.g. Becker & Marin 2009; Popper & Tuohy 2010; Wodniok et al. 2011; de Vries et al. 2017b: phenylpropanoid biosynthesis; Morozov et al. 2018: evolution of trans-acting small interfering RNAs). In their streptophyte/embryophyte versions they can perhaps be thought of as preaptions/exaptions (e.g. Becker & Marin 2009; Delaux et al. 2015; de Vries et al. 2018). Lang et al. (2010) noted that a considerable number of transcription-associated protein families evolved in the basal land plants or their immediate aquatic ancestors, many more than subsequently, but exactly where the later changes occur is unclear since nothing between Selaginella and angiosperms was included in this study. Boot et al. (2012) found that there was polar auxin transport in Chara; it is known from the sporophytic generation in mosses, but not liverworts (see also Harrison 2017b fig). The evolution of gibberellin receptors is of considerable interest, with major groups differing in receptor type (Yoshida et al. 2018), while the process of hydrolysis of inactive auxin conjugates (e.g. with amino acids) may be particularly distinctive in tracheophytes, but with links to what goes on in bryophytes and yet more basal streptophytes (Campanella et al. 2018).

There have been extensive changes to features involved in cell division, both mitosis and meiosis, in the basal streptophytes and also embryophytes. These have been evident for getting on to 50 years, and the cytological variation that was being discovered then had two major implications. It suggested first, that the old group of green algae should be broken up, and second, that a subset of these algae, the charophytes, had a number of similarities with land plants (Stewart & Mattox 1975). Thus the male gametes of bryophytes had asymmetrical sperm, while the insertion of of the flagellae onto the basal multilayered band of microtubules was sightly lateral, a number of streptophytes have phragmoplasts, and so on (Stewart & Mattox 1975; see also Mattox & Stewart 1984; Simon et al. 2006; Brown & Lemmon 2007 and references). Stewart and Mattox (1975) noted that the result of their synthesis was almost "a cytological classification", but its outlines have held. For characters of streptophytes and some of their subgroups, see e.g. Stewart and Mattox (1975), Leliaert et al. (2012: esp. Fig 4).

De Vries et al. (2016) note numerous features common in/characteristic of embryophytes that are also to be found in streptophytes, and Buschmann and Zachgo (2016) look at the evolution of the distinctive embryophyte mode of cell division from a similar point of view. See also de Vries et al. (2018: stress-signaling genes), Gao et al. (2018: R - disease resistance genes). The photorespiratory glycolate pathway occurs in peroxisomes rather than in mitochondria (Stabenau & Winkler 2005). The BIP multigene family is prominent (Friedl & Rybalka 2012). The duplication yielding the important GAPA/B gene pair occurred around here, and streptophytes have a particular isoform of glyceraldehyde-3-phosphate dehydogenase, GAPDH B (Peterson et al. 2006). Buschmann and Zachgo (2016) discuss the evolution of cell division. In basal streptophytes there is a cleavage furrow and the new transverse wall develops centripetally while in embryophytes the cell plate develops centrifugally where the preprophase band (p.p.b.) of microtubules had been. Basal streptophytes have a centrosome from which radiating microtubules develop and are involved in pulling the dividing chromosomes apart while in most embryophytes neither centrosome nor radiating microtubules develop. Changes in cell division are complex, and many are evident in various groups of streptophytes close to the embryophytes, although some show parallelism and even loss there (Buschmann & Zachgo 2016). Thus Mougeotia, in the clade sister to Embryopsida, lacks centrioles and has a p.p.b., while Coleochaete (same clade, or next clade down) has monoplastidic meiosis and centrioles (e.g. Brown & Lemmon 1993, 2011b). See also immediately below...

As is clear from the discussion on phylogeny below, clarifying the relationships of and within Zygnematales is clearly critical if we are to understand the evolution of land plants. However, whatever the sister-group relationships of embryophytes might be, Chara et al. are no longer likely candidates, and this questions an evolutionary scenario involving the evolution of ever more complex plant bodies.

Ecology & Physiology. It is likely that embryophytes are not the first streptophyte clade to live on land, as Stebbins and Hill (1980) suggested some time ago, indeed, they thought that Charophyceae s. str. might even have moved back to water (see also Harholt et al. 2016); if so, the latter would be unlikely to represent the morphology of the immediate ancestor of land plants, as had been thought in the preceding century. Klebsormidiophyceae, for example, can be terrestrial, and there are probable parallels between the problems they face and those faced by early terrestrial embryophytes - these centre on dessication, acquisition of nutrients, movement of water through the plant once the plant body reached any size, protection from ultraviolet (UV-B) radiation, photosythesis under high CO2 concentrations, and so on (Hori et al. 2014: e.g. several hormones and their receptors there, protection against light). Thus Pierangelini et al. (2018, see also 2017) noted that some Klebsormidiophyceae they studied were dessication tolerant (see also Holzinger et al. 2014) and avoided direct insolation by growing in low-light regimes, others could acclimate to high-light regimes but were not dessication tolerant (some of the responses/behaviours they studied were facultative). Moreover, dessication tolerance and photoacclimation were linked to the ability of taxa like Streptosarcina arenaria to form clumps ("packets") of cells, and in general cell aggregation, whether cell clumps or even biofilms, helped prevent dehydration (Pierangelini et al. 2018), indeed, Stebbins and Hill (1980) had noted that such three-dimensional growth forms would be resistant to dessication. Similarly, Coleochaete orbicularis, normally aquatic and with an orbicular, unistratose plant body, can grow on agar or quartz sand, and then it produces clumps of cells, some with thickened, layered walls, and with least some dessication tolerance (Graham et al. 2012), while in Zygnema dessication tolerance was evident in lipid-rich and thick-walled cells (Herberger et al. 2015: the preakinete stage, cells still in filaments). For the synthesis of the cell wall polysaccharides so important in land plants, but also found in more basal streptophytes, see Mikkelsen et al. (2014). Del-Bem (2018) noted that genes involved in xyloglucan metabolism are found in many streptophytes, although few in Chlorkybus and Mesostigma. He notes that xyloglucans are the most important component of the non-cellulosic non-pectic matrix of the primary cell wall, and that xyloglucan-producing streptophytes may have been components of early (pre-embryophyte) soil crusts, perhaps also helping in the aggregation of soil particles around algal cells (Del-Bem 2018); Raven (2018) discussed early soil formation; he thought that this might (p. 1139, his emphasis) have occured in the Proterozoic, yielding deeply entrenched river channels, however, mudstones, reflecting silicate weathering by roots, was evident only in the Late Ordovician. Indeed, although the cell wall of embryophytes is a distinctive structure, Harholt et al. (2016) noted that few of its components were unique to embryophytes; it likely represented the result of the adaptation of earlier streptophytes to life on land and its presence in Characeae s. str. represents the retention of ancestral (terrestrial) features. Taxa like Zygonema, Klebsormidium, etc., do not produce zoospores, again, a possible adaptation to life on land.

Interestingly, mosses and liverworts as well as at least some streptophytes immediately basal to embryophytes can take up mono- and/or disaccharides - glucose, fructose and sucrose - and thus mixotrophy could be the basal condition for embryophytes (Graham et al. 2010a, 2010b).

Genes & Genomes. Leliaert et al. (2012) summarize variation in the plastid and mitochondrial genomes in streptophytes and compare it with that in embryophytes, and some of this may convert to apomorphies when comparative data improve. For the evolution of the chloroplast genome, see de Vries et al. (2016); de Vries et al. (2017a) found an ultraconserved motif of the YCF1 gene in the higher streptophytes that is conserved across most land plants (see also Nakai 2015). The tufA gene has moved to the nucleus, although it is present, if very odd, in Coleochaete chloroplasts (e.g. Baldauf et al. 1990; de Vries et al. 2016). For the chloroplasts of Anthoceros and cycads (C.-S. Wu & Chaw 2015); the chloroplast ultrastructure of Coleochaete is in some ways similar to that of hornworts (Vaughn et al. 1992). A possible synapomorphy for the node that includes Mesostigma viride and the rest is the presence of the ndh and rps15 and the loss of the rps9 chloroplast genes, although Chara has the rps9 but not the rps15 genes (Martín & Sabater 2010); the presence of introns in the ndhA and ndhB genes may also be apomorphies around here, and Ruhlman et al. (2015) summarize the increase in complexity of the NADH dehydrogenase-like (NDH) complex in these streptophytes. Both Zygnema and the desmid Staurastrum secondarily lack the chloroplast inverted repeat, present in other streptophytes (Turmel et al. 2007), indeed, evidence now suggests that the IR may have been lost several times in the [Zygnematales + Desmidales] clade as well as in Coleochaete (Civán et al. 2014; Lemieux et al. 2016). In general the chloroplast genome in the [Zygnematales + Desmidales] is very labile, group II introns frequently being lost, and some phage/viral genes have been gained, although overall the genes it contains are much less labile (see esp. Lemieux et al. 2016). For similarities in the organisation of the chloroplast genome in particular between basal streptophytes and basal green algae, see Leliaert et al. (2016). See also discussion under Embryopsida below.

Phylogeny. Resolving the relationship of the polyphyletic prasinophytes, mostly Chlorophyta, has been important (e.g. Lewis & McCourt 2004; Niklas & Kutschera 2010; esp. Leliaert et al. 2016 and references). Mesostigma viride, the only streptophyte with an eye spot, used to be included in that group, but along with Chlorokybus it is sister to the other streptophytes, as is indicated by nuclear and some, but by no means all, organellar genes (E. Kim et al. 2006: isoprenoid synthesis pathways and glycolate oxidizing enzymes agree). Its zoospores have scales but lack a cellulose cell wall (J. Petersen et al. 2006) and it links with the non-motile Chlorkybus (Simon et al. 2006); these two genera are placed outside a [Chlorophyta + Streptophyta] clade by Gitzendanner et al. (2018).

It was often thought that Characeae/Charales s. str. (inc. Nitella, but not much else) were the immediate sister group of land plants (e.g. Graham 1993: still useful and readable; Karol et al. 2001; Turmel et al. 2003; Delwiche et al. 2004; Qiu et al. 2007: quite strong support), partly because they seemed to be intermediate in a progression between simple "algal"-like morphologies and the more complex land plants - they are filamentous, growth is by an apical cell, and they are oogamous. However, for twenty years or so molecular evidence has suggested otherwise (e.g. Hedderson et al. 1998: small subunit rRNA). Turmel et al. (2006, 2007) found Zygnematales to occupy this position in a number of analyses based on complete chloroplast genome sequences - a commonly-found set of relationships was [Chara [Chaetosphaeridium [[Zygnema + Staurastrum] + Embryopsida]]] (see also Chang & Graham 2011: Staurastrum not included). Similar relationships are commonly recovered, e.g. [Nitella [[Spirogyra, Closterium, Chaetosphaeridium, etc.] [Coleochaete + Embryopsida]]] (Finet et al. 2010), while B. Zhong et al. (2013a, b; Zhong et al. 2014a) found the variant relationships [Charales [Coleochaetales (inc. Chaetosphaeridium) [[Zygnematales + Desmidales] + Embryopsida]]] or [[Coleochaetales + Zygnematales] Embryopsida]; see also the relationships in Simon et al. (2006) and Sayou et al. (2014). Springer and Gatesy (2014) reanalysed the data of Zhong et al. (2013a) and found different topologies depend on how the data were analysed, but Charales were never sister to embryophytes. Overall, the relationships recovered around here are most often [[Chlorokybus + Mesostigma] [Klebsormidium [Nitella [[Coleochaete + Chaetosphaeridium] [[Penium + Spirogyra] + embryophytes]]]]], e.g. as found by Laurin-Lemay et al. (2012: q.v. for details) in a reanalysis of the data used by Finet et al. (2010) i.a. excluding contaminated data such as rotifer and diatom sequences; see also Cooper (2014) and X.-X. Shen et al. (2017: evaluation of support).

Evidence for the sister group relationships of embryophytes and Spirogyra and its relatives is now quite strong (Leliaert et al. 2012: literature; Timme et al. 2012; see also Wodniok et al. 2011; Ruhfel et al. 2014; Wickett et al. 2014; Puttick et al. 2018: two transcriptome analyses; Davis et al. 2014a; Civán et al. 2014; Lemieux et al. 2016 - all chloroplast genomes). The [Penium + Spirogyra] clade has a number of apomorphies, including the loss of cilia and also the loss of morphological complexity (Wodniok et al. 2011; Timme et al. 2012). For divergence within Coleochaete and its morphological variation, see Delwiche et al. (2002). Many genera of Desmidiaceae are polyphyletic (Friedl & Rybalka 2012 and references).

Classification. For a classification of life, see Ruggiero et al. (2015).

EMBRYOPSIDA Pirani & Prado

Gametophyte dominant, independent, multicellular, initially ±globular, not motile, branched; showing gravitropism; glycolate oxidase +, glycolate metabolism in leaf peroxisomes [glyoxysomes], acquisition of phenylalanine lysase* [PAL], flavonoid synthesis*, microbial terpene synthase-like genes +, triterpenoids produced by CYP716 enzymes, CYP73 and phenylpropanoid metabolism [development of phenolic network], xyloglucans in primary cell wall, side chains charged; plant poikilohydrous [protoplasm dessication tolerant], ectohydrous [free water outside plant physiologically important]; thalloid, leafy, with single-celled apical meristem, tissues little differentiated, rhizoids +, unicellular; chloroplasts several per cell, pyrenoids 0; centrioles/centrosomes in vegetative cells 0, microtubules with γ-tubulin along their lengths [?here], interphase microtubules form hoop-like system; metaphase spindle anastral, predictive preprophase band + [with microtubules and F-actin; where new cell wall will form], phragmoplast + [cell wall deposition centrifugal, from around the anaphase spindle], plasmodesmata +; antheridia and archegonia +, jacketed*, surficial; blepharoplast +, centrioles develop de novo, bicentriole pair coaxial, separate at midpoint, centrioles rotate, associated with basal bodies of cilia, multilayered structure + [4 layers: L1, L4, tubules; L2, L3, short vertical lamellae] (0), spline + [tubules from L1 encircling spermatid], basal body 200-250 nm long, associated with amorphous electron-dense material, microtubules in basal end lacking symmetry, stellate array of filaments in transition zone extended, axonemal cap 0 [microtubules disorganized at apex of cilium]; male gametes [spermatozoids] with a left-handed coil, cilia 2, lateral, asymmetrical; oogamy; sporophyte +*, multicellular, growth 3-dimensional*, cuticle +*, plane of first cell division transverse [with respect to long axis of archegonium/embryo sac], sporangium and upper part of seta developing from epibasal cell [towards the archegonial neck, exoscopic], with at least transient apical cell [?level], initially surrounded by and dependent on gametophyte, placental transfer cells +, in both sporophyte and gametophyte, wall ingrowths develop early; suspensor/foot +, cells at foot tip somewhat haustorial; sporangium +, single, terminal, dehiscence longitudinal; meiosis sporic, monoplastidic, MTOC [= MicroTubule Organizing Centre] associated with plastid, sporocytes 4-lobed, cytokinesis simultaneous, preceding nuclear division, quadripolar microtubule system +; >1000 spores/sporangium, sporopollenin in the spore wall* laid down in association with trilamellar layers [white-line centred lamellae; tripartite lamellae]; plastid transmission maternal; nuclear genome [1C] <1.4 pg, main telomere sequence motif TTTAGGG, KNOX1 and KNOX2 [duplication] and LEAFY genes present, ethylene involved in cell elongation; chloroplast genome with introns (not: Mesostigma), close association between trnLUAA and trnFGAA genes [precursors for starch synthesis], tufA, minD, minE genes moved to nucleus; mitochondrial trnS(gcu) and trnN(guu) genes +. Extant: 7 clades, 403,911 spp. (Nic Lughadha et al. 2016).

Note: In all node characterizations, boldface denotes a possible apomorphy, (....) denotes a feature the exact status of which in the clade is uncertain, [....] includes explanatory material; other text lists features found pretty much throughout the clade. Note that the precise node to which many characters, particularly the more cryptic ones, should be assigned is unclear. This is partly because homoplasy is very common, in addition, basic information for all too many characters is very incomplete, frequently coming from taxa well embedded in the clade of interest and so making the position of any putative apomorphy uncertain. Then there are the not-so-trivial issues of how character states are delimited and ancestral states are reconstructed (see above).

Age. Clarke et al. (2011: other estimates, c.f. topology) suggested an age for crown Embryopsida of (815-)670(-568) Ma, Cooper et al. (2012) estimated its age at (519-)493(-469) Ma, and Magallón et al. (2013: including temporal constraints) an age of around (480.4-)475.3-474.6(-471) Ma (see the constraint age in Heinrichs et al. 2007) and a stem age of around (962.5-)913, 911(-870) My; the lowest crown date is around 439 Ma (Magallón & Hilu 2009). An age of (629.5-)530(-449) Ma is suggested in B. Zhong et al. (2014b: fossil calibration), ca 487 Ma in Evkaikina et al. (2017), 515-473.6 Ma in Morris et al. (2018: HPD, other dates and extensive discussion) and (485-)482(-473) Ma (Lutzoni et al. 2018: N.B., the stem could be up to 727 My); see also P. Soltis et al. (2002) and Guindon (2018), a variety of ages that depend on calibrations.

455-454 Ma fossils of leaves with distinctive cells remarkably similar to those of extant Sphagnum from Ordovician deposits in Wisconsin are the earliest known vegetative remains of an embryophyte (Cardona-Correa et al. 2016).

Includes Bryophyta s.l., Tracheophyta. Note that all groups below are extant crown groups unless specified otherwise.

Evolution: Divergence & Distribution. >Laenen et al. (2014) give ages and diversification rates for well over 150 major clades throughout land plants.

Many of the bolded characters in the characterization above are apomorphies of clades of subsets of streptophytes + embryophytes along the lineage leading to the embryophytes, not apomorphies of crown-group embryophytes per se; these latter are indicated by asterisks (see also Timme et al. 2012: immediate relatives of embryophytes; Bowman et al. 2017 for apomorphies). Of course, all the features of sporophytes mentioned could be dignified by asterisks... Characters supporting relationships between Embryopsida and subsets of streptophytes include many details of cell division, occurrence of an apical cell in the gametophyte (perhaps), numbers and types of introns in the chloroplast DNA, cilium ultrastructure, occurrence of sporopollenin, retention of the zygote on the haploid plant, and nrDNA in a single array; see also above.

Embryophytic plants show alternation of generations, a sporophytic generation being interpolated into a life cycle that, with the exception of the diploid zygote, was entirely haploid/gametophytic, i.e. similar to the life cycle of some of the more basal aquatic streptophyes. It is unlikely that these alternating generations are the result of the divergence of initially morphologically similar diploid sporophytic and haploid gametophytic generations (e.g. Haig 2008, 2015; Gerrienne & Gonez 2011). The BELL1 homeobox gene may be involved in the interpolation of mitotic cell divisions between zygote formation and meiosis in early land plants (Horst & Reski 2016), although the gene is also known from Chlamydomonas, not at all closely related... The larger and more complex plant body would allow the production of meiospores, the immediate products of meiosis, in larger numbers (e.g. Brown & Lemmon 2011a; Qiu et al. 2012; Edwards et al. 2014). This change may have been driven by the unpredictability of fertilization (Haig 2015). (Interestingly, in Coleochaete the large, maternally-provisioned zygote may produce up to 32 zoospores, although exactly when meiosis occurs is unclear - Haig 2015.)

Thus there may have been an initial development of a diploid spore-producing sporangium (= sporogonium) borne on the gametophyte, with the later elaboration of a full-blown free-living sporophyte in the polysporangiophyta (e.g. Kaplan 1997: chap. 19; Kato & Akiyama 2005; Qiu et al. 2012: extensive discussion; Ligrone et al. 2012b), the sporophyte becoming independent of the parental gametophyte (Haig 2015; Qiu et al. 2012). Some streptophytes and nearly all land plants have placental transfer cells or their equivalents on the sporophytic and/or gametophytic sides of where the two generations join; these cells have labyrinthine wall ingrowths that mediate an initial transfer of nutrients from the gametophyte to the sporophyte (see Gunning & Pate 1969b; Ligrone et al. 1993: much detail; Hilger et al. 2002; Carapa et al. 2003; Vaughn & Bowling 2008; Renzaglia & Whittier 2013 for a tree). However, we lack reliable knowledge of life cycles in most charophyte algae, and this hampers our understanding of the events that led to the development of the alternation of generations in land plants (Haig 2010, 2015). For studiers of the early evolution of land plants with apomorphies at all levels and incorporating fossil members, see e.g. Kenrick and Crane (1997; also Kenrick et al. 2004; Doyle 2013; etc.), however, characters now need to be re-evaluated and pegged to their appropriate nodes, especially given the probability that bryophytes s.l. are monophyletic and sister to vascular plants.

Yue et al. (2012) suggested that numerous genes, including those involved in many embryophyte functions like cuticle and lateral root formation, had been acquired by lateral transfer from fungi and bacteria (see also Emiliani et al. 2009; R. Chen et al. 2019: glycoside hydrolase subfamily GH5_11, from fungi, esp. cellulose degradation). Along the same lines, an actinoporin in Physcomitrella patens, Tortula ruralis and Selaginella lepidophylla is involved in water stress responses, overexpression in the first species making the gametophore tolerant to dehydration (Hoang et al. 2009). Similar actinoporins are known from sea anemones, and Hoang et al. (2009) suggested that lateral gene transfer from sea anemones to land plants had occurred, and that bryoporin, the actinoporin in land plants, was involved in the response to water stress in the common ancestor of these three plant, i.e. the common ancestor of extant land plants.

In mosses, at least, the great majority - ca 95% - of genes are expressed in both sporophyte and gametophyte generations (Szövényi et al. 2010; c.f. in part O'Donoghue et al. 2013). This is in line with the proposal by Banks et al. (2011) that some of the genes involved in the patterning and differentiation of vascular tissue were present in the ancestral (gametophyte dominant) land plant and were recruited by the sporophytic generation. Indeed, in the fern Polypodium ca 97% of the genes are expressed in both generations, only 10% of the genes differing in transcription levels between the two (Gigel et al. 2016). Frank and Scanlon (2014) proposed a model for sporophyte evolution that is similar to that for vascular tissue evolution. However, there are often substantial differences between the two generations in the expression of genes controlling apical meristem growth and auxin polarity (e.g. Fujita et al. 2008; Sakakibara et al. 2008; Harrison 2017a, b for literature), and other studies found that over 12% of the transcriptome of the moss Physcomitrella switched in the transition from gametophyte to sporophyte, especially those genes involved in carbohydrate and energy metabolism (O'Donoghue et al. 2013; see also Szovenyi et al. 2010). It is only in angiosperms that a substantial proportion (ca 25%) of genes are expressed in the sporophyte alone (Szövényi et al. 2010). Interestingly, in both angiosperms and Physcomitrella genes expressed in the gametophyte were younger than those expressed in the sporophyte (O'Donoghue et al. 2013; Cui et al. 2015; Gossmann et al. 2016). Floyd and Bowman (2007) discuss developmental changes possibly occurring at this node, and Friedman et al. (2004) the evolution of plant development.

With a monophyletic bryophytes s.l. how we think of land plant evolution is changing, and as Puttick et al. (2018: p. 10) noted, "Marchantia and other liverworts might not serve as an appropriate model for anything other than liverworts themselves." (c.f. e.g. Shimamura 2016; Bowman et al. 2017). The exact relationship between the sporophytes of polysporangiophytes and of bryophytes s.l. is unclear (e.g. Shaw et al. 2011), especially given the uncertainties in the phylogeny. The stalk of a moss capsule may not be strictly homologous to the branching sporophyte axis of the polysporangiophytes (Kato & Akiyama 2005; Qiu et al. 2012). Sterilization of the sporangial axis or simple elaboration of a bryophyte sporangium leading to the evolution of a branched sporophyte are ideas that have been suggested relating bryophytes and polysporangiophytes, but prolongation of embryonic growth is perhaps more likely (Tomescu et al. 2014). Frank and Scanlon (2014) found that there was expression of meiosis gene families early in the development of both Physcomitrella (see also O'Donoghue et al. 2013: carbohydrate metabolism genes) and Marchantia sporophytes, which would perhaps make their extended growth unlikely, however, in the stem apex of maize and in the gametophyte apex of mosses, but not liverworts, similar patterning gene families were upregulated (see also Fujita et al. 2008; Sakakibara et al. 2008). A scenario for the evolution of a complex sporophyte may involve delayed expression of genes involved in meiosis, which would allow for indeterminate development of a sporophyte in which these patterning genes were expressed (Frank & Scanlon 2014). Genes expressed in the gametophyte may be essential for the subsequent elaboration of the sporophyte. Interestingly, disruption of PIN [auxin efflux facilitator] in e.g. Physcomitrella can lead to branching of the sporophyte (e.g. Bennett et al. 2014).

Some sporophytes of early polysporangiophytes were very small. Those of the Silurian cooksonioids in particular are 1 mm or substantially less across and were probably dependent physiologically on the gametophyte (Rothwell 1995; Boyce 2008b; see also Edwards et al. 2014). However the ca 432 Ma Cooksonia barrandei from the early Silurian of the Czech Republic has a dichotomously-branched plant body over 5 cm long and 1-2.5 mm or more across and is likely to have been photosynthetic (Libertín et al. 2018), although Cooksonia-type fossils from Ireland around 427 Ma are notably smaller (see Edwards et al. 1983) and perhaps unable to photosynthesize. Conversely, some early gametophytes were quite elaborate structures, although different in morphology from the sporophytes, and they even had conducting tissue, stomata, a cuticle, and were apparently homoiohydric - i.e. they had many features now associated with the sporophyte (Remy & Haas 1991; Edwards 1993; Edwards et al. 1998; Edwards & Richardson 2004; Taylor et al. 2005; Gerrienne & Gonez 2011; Ligrone et al. 2012a). (Cooksonia does have vascular tissue in the sporophyte, although apparently it is not found in some species - see Harrison 2017a, b and literature.) A stage in the evolution of vascular plants may involve an ancestor in which the generations were pretty much isomorphic, perhaps similar to some of the plants from the Lower Devonian Rhynie Chert of some 410 Ma, and they figure largely in attempts to understand the evolution of land plant life cycles (Niklas & Kutschera 2009, 2010). Some early vascular plants had multicellular stem apices (Hueber 1992). The sequence may be as follows: Gametophyte only free living, sporophyte unbranched, dependent on the gametophyte → gametophyte only free living, sporophyte branched, dependent on the gametophyte → generations ± isomorphic, both free living → sporophyte dominant, gametophyte free-living → sporophyte dominant, gametophyte dependent on sporophyte (see also Tomescu et al. 2014; c.f. Qiu et al. 2012). A modified homologous theory of alternation of generations in which the two generations were at one stage initially more or less similar, and "the gametophyte body plan was substantially expressed in the sporophyte" (Kenrick 2017: p. 8) is on the cards (see also Libertín et al. 2018). The gametophytic generation became transformed - either ultimately much reduced, as in extant vascular plants, or remaining/becoming quite elaborate, as in extant bryophytes s.l. - is on the cards (Kenrick 2017; Libertín et al. 2018).

A variety of spores, including permanent tetrads, may have been produced by protoembryophytic plants whose whole sporophyte was little more that these spores. Cryptospores, initially largely defined by what they were not (Strother 1991; Volkova et al. 2016 for meiosis), are known from the mid-Cambrian onwards, and cryptospores that are like bryophyte spores appear in the middle Ordovician, about 40 Ma later and ca 476 Ma (e.g. Gray 1993; Wellman 2004a; Rubinstein et al. 2010: 473-471 Ma spores from Argentina; Brown & Lemmon 2011a; Gerienne et al. 2016 and references; for the evolution of the spore walls of land plants, see also Blackmore & Barnes 1987). Fragments of plant bodies of very early land plants appear in Late Ordovician rocks from Oman (Wellman et al. 2003). Edwards et al. (2014) discuss what they call cryptophytes, plants that cannot be confidently assigned to any extant group, from this period, and the cryptospores they produced, which were hilate monads, dyads, and permanent tetrads. However, most literature suggests that the earliest fossils are of bryophytes of some sort, probably liverworts, vascular plants being souped-up bryophytes (e.g. Kenrick 2000; Wellman et al. 2003; Rubinstein et al. 2010) - at least, they are in some tellings of the tale! Fossils remarkably like Sphagnum have been found in Ordovician rocks ca 455 Ma in North America (Graham et al. 2013, esp. Cardona-Correa et al. 2016).

Salamon et al. (2018) describe possible polysporangiate fossils in Late Ordovician rocks ca 445 Ma from Poland. Spores there have irregular trilete markings, and hilate/trilete spores that do not remain as tetrads appear in the Early Silurian ca 443 Ma and may mark the origin of vascular plants (Steemans et al. 2009; Kenrick et al. 2012; Gerienne et al. 2016 and references), although not only vascular plants have such spores (Salamon et al. 2018 and references). A better understanding of early fossil polysporangiophytes is particularly important.

For apomorphies of bryophyte groups and the early polysporangiophytes, see Garbary and Renzaglia (1998), Ligrone et al. (2012a). Gerienne et al. (2016) characterize many embryophyte clades up to and including spermatophytes (but note that details of the topologies of both trees in their Fig. 2 differ from the topology of the relationships accepted here). Characters are placed on a land-plant tree in Z.-H. Chen et al. (2017: Fig. 1), but several are in odd places, c.f. also the tree topology. These, too, all have to be worked through.

There is a useful summary of the evolution of cell division in Buschmann and Zachgo (2016). Brown and Lemmon (e.g. 1990, 2007; Brown et al. 2010) and others have unravelled some of the complexity of both mitotic and meiotic cell divisions in land plants, particularly extreme in liverworts; for meiosis, see also Renzaglia and co-workers (e.g. Renzaglia et al. 2000a, Renzaglia & Garbary 2001). De Vries and Gould (2017) discuss the evolution of polyplastidy - tracheophytes are consistently polyplastidic, although there is variation in bryophytes s.l. - and movement of the minD and minE genes out of the chloroplast, movement that is thought to be necessary for the evolution of this condition, occured in the basal part of the streptophyte clade where there is also much variation in chloroplast number. This may be correlated with patterns of microspore division (Rudall & Bateman 2007).

Although centrosomes and centrioles, which show vertical inheritance, are lost (Murata et al. 2007), similar structures (but they are not inherited) may develop from the blepharoplast during meiosis from microtubule organizing centres (MTOCs), and these vary considerably (Southworth & Cresti 1998; Brown & Lemmon 2007). Indeed, in Marchantia polymorpha the nature of the MTOC changes during meiosis (Brown et al. 2007). In general, γtubulin, involved in the nucleation of microtubules, is highly migratory. Brown and Lemmon (1997, see also 2008) reviewed the distribution of the quadripolar microtubule system in land plants; it occurs in all three bryophyte groups, in lycophytes, and in Marattiales, at least. This system is linked with monoplastidy, the MTOC being associated with the plastid. In streptophytes as a whole basal bodies develop de novo immediately prior to the formation of motile sperm cells (Wastenays 2002). Interestingly, the preprophasic band (p.p.b.) is sometimes absent, as in the protonema of at least some mosses, and even in megaspore development and in the endosperm of flowering plants, although elements of the p.p.b. apparatus may still be present (Webb & Gunning 1991; Buschmann & Zachgo 2016 and references).

Members of Klebsormidiales, Coleochaetales, Desmidaceae and Zygnematales have pyrenoids, and so the loss of pyrenoids is likely to be an apomorphy for embryophytes (Cook 2004a, b). Hodges et al. (2012) survey the evolution of land plant cilia, in part focussing on the ciliome and the role it continues to play in land plant gametes even after the cilia themselves are lost.

See Mishler and Churchill (1984, 1985) for important early morphological phylogenetic analyses; Graham (1993) and Finet et al. (2010), both general; Graham et al. (2000), body plan; Stabenau & Winkler (2005), glycolate metabolism, from mitochondria to cytoplasmic; Becker & Marin 2009, Domozych et al. 2010, Popper & Tuohy 2010, Sørensen et al. (2010), distinctive cell wall polysaccharide polymers evolved in charophyceans or before; Comparot-Moss and Denyer (2009), evolution of starch synthesis. For other possible apomorphies, see e.g. Goffinet (2000), Renzaglia et al. (2000), H. Schneider et al. (2002: much useful information), Rensing et al. (2007a), Johnson and Renzaglia (2009), Doyle (2013) and Bowman et al. (2017). Carothers and Duckett (1982: bryophytes) and Renzaglia and Garbary (2001) discussed the evolution of the male gametes in land plants in some detail. For information on bryophyte s.l. body form easily placed in a phylogenetic context, see Goffinet and Buck (2013).

Ecology & Physiology. Although some of the problems of life on land faced by embryophytes may have been largely solved by their streptophyte ancestors (see below), such problems continued to shape the evolution of both gametophyte and sporophyte (e.g. Watkins et al. 2007a, b; Del Bem & Vincentz 2010; McAdam & Brodribb 2011; Willis & McElwain 2014; Graham et al. 2014; Proctor 2014; Raven & Edwards 2014; Robinson & Waterman 2014). Bateman et al. (1998), Hemsley and Poole (2004) and others discuss the physiology and ecology of early land plants.

Mycorrhizal associations of plants with fungi, perhaps particularly Mucoromycotina, may have made it possible for early plants to establish themselves on land (e.g. B. Wang et al. 2010; Selosse et al. 2015; Feijen et al. 2017; Kamel et al. 2018). Thus Pirozynski and Malloch (1975: p. 162), who emphasized the importance of a close association between fungi and early land plants, and indeed throughout land plant history, noted that "land plants never had any independence, for if they had, they could never have colonized the land". There are in fact a variety of scenarios for the origin and evolution of mycorrhizal associations (Field & Pressel 2018 for a summary) depending i.a. on the ages of the various protagonists. For more discussion, and also suggested ages of this association, see below. Mitchell et al. (2016) suggested that soil formation in current lichen-liverwort communities in Iceland with their associated mycorrhizal fungi, cyanobacteria, etc., might be an analogue for conditions in Rhynie chert communities; the soil formed was shallow, ca 1.5 cm deep, but it included smectite-type clay minerals, unlike the deeper (ca 3.5 cm) soils largely made up of mineral grains that developed in moss communities. Raven (2018) noted the role that glomalins, proteins in the walls of the glomeromycotes that are involved in arbuscular mycorrhizal symbioses, might play both in soil aggregation and carbon sequestration.

Many changes evident in land plants are connected with the development of a distinctive and complex phenolic metabolism, the phenylpropanoid pathway (Vogt 2010; de Vries et al. 2017b)/phenolic network (Renault et al. 2017). The products of this network include sporopollenin, lignins, integral to the support and water-conducting facilities of vascular plants in particular, suberin, flavonoids, involved in UV screens, and they are antioxidants, pigments, etc. (Renault et al. 2017). Phenylalanine lysase (PAL) is the first step in this pathway. Emiliani et al. (2009) suggested that embryophytes acquired this gene via lateral transfer from a bacterium; they found that it was also present in fungi, but absent from Nostoc. However, Delwiche et al. (1989) recorded lignin-like compounds (positive Maule reaction) from Coleochaete and Ligrone et al. (2008) lignin-related compounds from Nitella, so exactly where PAL moved into the streptophyte clade is uncertain. Some of these compounds were synthesised by streptophytes, especially by Mesostigma viride, but not by immediately subsequent clades on the tree (de Vries et al. 2017b).

The second step in the synthesis of phenolic compounds is mediated by the cytochrome P450, the CYP73 gene, a cinnamate hydroxylase, and this is absent in streptophytes but duplicated in seed plants/?angiosperms in particular (Renault et al. 2017). Much of the pathway by which lignin is synthesized in vascular plants is also found in mosses (Gómez-Ros et al. 2007; Z. Xu et al. 2009, see also Guo et al. 2010), but in some cases pathways in mosses and liverworts may differ (see below). S-lignin, made up of syringyl units, has been found in some liverworts and is scattered elsewhere in vascular plants, including Selaginella, and it is very common in flowering plants (Z. Xu et al. 2008; Li & Chapple 2010; Espiñeira et al. 2010; see also Gómez-Ros et al. 2007). For further discussion about the distribution of lignin precursors, see Labeeuw et al. (2015) and Chao et al. (2017).

One of the major changes, or better, complex of changes, that facilitated the spread of land plants may have been the evolution of sporopollenin-covered spores, and sporopollenin is known from several streptophytes including Coleochaete (Delwiche et al. 1989). Here sporopollenin is associated with the inner wall of the zygote (Gray 1993; Blackmore & Barnes 1987; see also Wallace et al. 2011; de Vries & Archibald 2018, etc.), and in land plants it may have become associated with the walls of the haploid spores by a complex process involving precocious initiation of cytokinesis (hence the often quadrilobed, quadripolar microtubule system), acceleration of meiosis, delay in wall deposition, etc., a mixture of heterochrony and heterotopy (Brown & Lemmon 2011a). Sporopollenin composition is remarkably stable in land plants, and the sporopollenin of the walls of lycophyte fossil spores ca 310 Ma, those of extant lycophytes, and in the walls of the embryos in some Charales is quite similar (Fraser et al. 2012), although the sporopollenin of conifers, at least, may differ. Homologues of just about all genes involved in pollen wall development in Arabidopsis are to be found in Selaginella and Physcomitrella, emphasising the fundamental similarity of pollen and spores in embryophytes (Wallace et al. 2011). Wellman (2004b) noted that sporopollenin deposition was associated with white line-centred lamellae.

Isoprenoids (xanthophylls and tocopherols) that protect against photo-oxidation and other insults of a dry, high light environment are widely distributed in land plants and in some of their aquatic relatives, although some carotenoids involved in these functions may have evolved in land plants (Esteban et al. 2009).

Field et al. (2015c) looked at the changing interactions between extant liverworts and both glomeromycotes and mucoromycotes in the context of the almost four-fold decrease in atmospheric CO2 concentrations from the Paleozoic to the present. Interestingly, if glomeromycotes were early associates with land plants it is notable that the efficiency of acquisition of both phosphorus and nitrogen by the liverwort from the fungus decreased with decreasing CO2 concentrations (and also when the liverwort had relationships with both kinds of fungi), the amount of carbon going to the fungus decreased, and overall plant growth increased. If mucoromycotes were the associates phosphorus and nitrogen transfer to the plant increased (see also Rimington et al. 2016). There are similarities in the microbiomes of streptophytes and bryophytes, in particular, similar nitrogen-fixing and methanotrophic organisms occur in both, and this has implications both for the evolution of the atmosphere and for the organisms involved (Knack et al. 2015).

The ability to tolerate dry conditions is obviously very important for land plants. Genes involved in dessication tolerance in the spores of the earliest land plants may have been coopted into dessication tolerance in gametophytes (Oliver et al. 2005). Plastids are important here, since the plant's perception of stress is associated with cellular cross-talk, from the plastid to the nucleus (retrograde signaling) and from the nucleus to the plastid (prograde signaling), the plastid being the site of many stress responses. It has been shown that Coleochaete and in particular Zygnema invest heavily in plastid-targeted proteins, while abscisic acid, central to stress responses in embryophytes, has also been found in Zygnema (de Vries et al. 2018; also Rippin et al. 2017; see Jarvis & López-Juez 2013 for plastids). Tolerance of extreme dessication is quite widespread in both mosses and liverworts (Proctor & Pence 2002), and even their antheridia are notably dessication-tolerant (Stark et al. 2016 and references). Dessication tolerance may even be the ancestral condition for land plants (Oliver et al. 2000), being lost e.g. in some mosses and liverworts, perhaps several times, and also in stem hornworts (and since regained several times: Oliver et al. 2005); it can be induced in some mosses and liverworts (Oliver et al. 2005; see also Proctor & Tuba 2002). The advantage of dessication tolerance in early embryophytes would seem self-evident, but given what we know of relationships and the vagaries of ancestral state reconstruction, pinning down exactly where changes occured is difficult. Interestingly, fern gametophytes are dessication-tolerant, and as Watkins et al. (2007a: p. 716) observed, "fern gametophytes are, for all intents and purposes, bryophytes". Late Embryogenesis Abundant (LEA) genes expressed in response to water stress - dessication in particular - are known from the moss Physcomitrella (Rensing et al. 2007a), Selaginella (van Buren et al. 2018a), Picea, and throughout flowering plants (Artur et al. 2018), and their progenitors are to be found in aquatic streptophytes (e.g. Wodniok et al. 2011; Artur et al. 2018 and references). Thus the oospores of Chara are dessication-tolerant (Oliver et al. 2005; Gaff & Oliver 2013) and desication tolerance in Klebsormidium is similar at the genic level to that in land plants (Holzinger et al. 2014) - see also above.

Interestingly, important elements of stress tolerance and retrograde signalling are similar in both streptophyte algae and in stomatal regulation in embryophytes in general, "suggesting that intricate cellular communication networks were already in place to prime stomatal regulation" (Chao et al. 2019: p. 5019). Indeed, understanding the relationship - both phylogenetic and functional - between stomata in bryophytes s.l. (mosses and hornworts) and those in vascular plants has always been difficult, but now understanding more about stomata in stem-group vascular plants is critical: Did stomata evolve independently in stem bryophytes s.l. and stem tracheophytes? Moreover, having "presence of stomata" as an apomorphy for extant embryophytes (the old stomatophytes included all embryophytes except liverworts) does not necessarily clarify the original context for their evolution (Puttick et al. 2018). There are not very many stomata on the capsules of either mosses or hornworts - they are at the base of the capsule in bryophytes and along the capsule in hornworts - and they seem not be involved in photosynthetic gas exchange, rather, they may facilitate the drying out of the capsule and hence aid in spore dispersal (Duckett et al. 2009; Pressel et al. 2011; Merced & Renzaglia 2013; Merced 2015; Chater et al. 2016, 2017; Brodribb & McAdam 2017; Merced & Renzaglia 2018). Renzaglia et al. (2017) and Pressel et al. (2018) discuss the role that stomata in hornworts play in the drying out of the capsule, etc., noting that the outer periclinal walls of the guard cells are pectic and very thin. The stomata, larger than epidermal cells here, unlike both mosses and vascular plants, open just once as the guard cells become more turgid (as with stomata in general, they have chloroplasts) and wall thickenings are laid down, and then they die and collapse, remaining open and allowing the drying of the contents of the sporangium (intercellular spaces here may be filled with liquid) (Merced & Renzaglia 2018). Stomata - sometimes called pseudostomata - of Sphagnum may also be involved in the drying out of the sporangium, opening as they lose turgor (Duckett et al. 2009). Although the guard cells to be found elsewhere in mosses are thickened and do not normally collapse, a function for stomata in the maturation of the capsule has been suggested in Physcomitrella, while stomata are probably involved in the drying and dehiscence of the capsule in other mosses (Renzaglia et al. 2017 and references). The Lower Devonian fossil Sporogonites and Tortilicaulis (relationships uncertain, perhaps polysporangiophytes, so the group from which Tracheophyta evolved) seem to have collapsing stomata similar to those of hornworts (Renzaglia et al. 2017). It has also been suggested that stomata may originally have increased transpiration and so improved the supply of nutrients to the sporangium (e.g. Edwards et al. 1998; Haig 2013).

Similar genes are involved in stomatal development in both vascular plants and some mosses (see also O'Donoghue et al. 2013; Vatén & Bergmann 2012; Sakakibara 2016; Chater et al. 2017; Merced & Renzaglia 2018); for the molecular basis of stomatal development, see also Peterson et al. (2010). However, whether the stomata of Sphagnum are like those of other mosses in such respects is unclear (Merced 2015). Stomata are separated by one or more epidermal cells, and for the literature on stomatal patterning, i.e. the development of stomata and their associated/surrounding cells, of extant and extinct embryophytes, see Rudall et al. (2013a), etc.. Some of the genes involved in stomatal development Physcomitrella retain their function when moved to Arabidopsis (Caine et al. 2016). However, as is clear from work on features like C4 photosynthesis, independent acquisitions of the "same" feature can involve remarkable similarities at the molecular level.

Stomatal behaviour in vascular plants and bryophytes s.l. may differ (McAdam & Brodribb 2011, esp. Fig. 4), for instance, abscisic acid plays a crucial role in control of stomatal opening only in seed plants, even if similar genes involved in abscisic acid metabolism are found throughout land plants (McAdam & Brodribb 2012; McAdam et al. 2016; c.f. Hõrak et al. 2017) and stomata in hornworts behave quite differently (Renzaglia et al. 2017, see above). Chater et al. (2011) suggested that stomata of the moss Physcomitrella responded to at least some environmental stimuli rather like those of flowering plants (O'Donoghue et al. 2013), abscisic acid being involved in both (see also Beerling & Franks 2009; Chater et al. 2014; references in Raven & Edwards 2014; Lind et al. 2015). Genes involved in the signalling pathway in guard cell opening/closing are known from liverworts (Chater et al. 2014; Lind et al. 2015), but McAdam et al. (2016) suggest that in bryophytes such genes are primarily involved in spore dormancy, and it does seem that bryophyte stomata function rather differently than those in vascular plants (Brodribb & McAdam 2017). Field et al. (2015b) found that stomatal density and aperture size in taxa from mosses and hornworts were largely unresponsive to changing CO2 concentrations. There is more discussion about stomata under extant tracheophytes; clearly, more work is neeeded here.

Recent ideas of relationships (bryophytes s.l. monophyletic, sister to vascular plants) and dates (crown-group land plants 515-473.6 My) need to be taken into account as one thinks of the evolution of the biosphere (Puttick et al. 2018; Morris et al. 2018). Raven and Edwards (2014) list estimates of the net photosynthetic rates for bryophytes and other early land plants. Back in the Late Ordovician ca 450 Ma and with an atmospheric CO2 concentration about eight times today's levels, land plants along with lichens may have supported a level of chemical weathering two to three times that of today's vegetation (Porada et al. 2016). Such high levels are likely to have been temporary as embryophytes spread, nutrients became limited and nutrient recycling in the developing soil/humus layers increased, but they may have helped precipitate the later Permo-Carboniferous glaciation (Porada et al. 2016). For photosynthesis in bryophytes and early embryophytes, see Graham et al. (2014) and other articles in Hanson and Rice (2014).

Y. Zhang et al. (2019) look at gravitropism in land plant roots and rhizoids, and although sampling is skimpy, the results suggest strongly that the rate of response to gravity changes in roots and rhizoids increased considerably at the extant seed plant node.

Plant-Animal Interactions. Salicylic acid, involved i.a. in resistance against pathogens that get their nutrition from living cells of the host and against some phloem feeding insects, and jasmonic acid, involved i.a. in resistance against parasites that get their nutrition from dead cells of their host and against chewing herbivores have an antagonistic relationship in many seed plants; genes involved in the relationship between the two appear to be found in all land plants (Thaler et al. 2012).

Some caterpillars of Micropterigidea, a jawed, lepidopteran clade that is perhaps Jurassic in age and sister to all other leps, are detritivores, but others eat mosses (e.g. Atrichum) and especially liverworts (Imada et al. 2011; Regier et al. 2015 and references; see also Hosts, consulted iii.2014), although they also eat angiosperms (Davis & Landry 2012 and references). On balance, evidence suggests that the Araucaria-eating jawed moths, Agathiphagidae, are sister to Micropterigidae (Regier et al. 2015; Kristiansen et al. 2015; c.f. Heikkilä et al. 2015; Mitter et al. 2016). For the host plants of other jawed moths, see also Nothofagaceae. Caterpillars of Mnesarchaeidae, members of a basal glossatan (= all other lepidoptera) clade, live in silk tunnels and eat mosses and liverworts, as well as fern sporangia, etc., their sister taxon, Hepialoidea, with many more species, are also concealed feeders, but they eat just about everything (Regier et al. 2015).

Bacterial/Fungal Associations. Associations between embryophytes and fungi, initially with the gametophytes of the former, were established very early in the Silurian/Devonian (Selosse & Tacon 1998; Redecker et al. 2000b; Nebel et al. 2004; Köttke & Nebel 2005; see also Strullu-Derrien et al. 2014b, 2017; Selosse et al. 2015; Rimington et al. 2017, etc.). Four main fungal groups, Glomeromycotina, Mucoromycotina, Ascomycota and Basidiomycota, are associated with embryophytes, the first two forming arbuscular mycorrhizal (AM) and the last two ectomycorrhizal (ECM) associations. Relationships between these fungi are probably [[Glomeromycotina [Mortierellomycotina + Mucoromycotina]] [Ascomycota + Basidiomycota]] (Spatafora et al. 2016). All four groups are associated with liverworts (Read et al. 2000; Duckett et al. 2006b; Pressel et al. 2010; Bidartondo et al. 2011; Field et al. 2015c, d; Weiß et al. 2016). Bidartondo et al. (2011; see also Pressel et al. 2010) found that Endogone-like fungi (Mucoromycotina) formed ECM associations with Treubia, Haplomitrium, some hornworts, etc., and surmised that this might represent the original land plant-fungus association (see also Rimington et al. 2014; Field et al. 2015d); Chang et al. (2018) dated the formation of this association to ca 420 Ma, mid-late Silurian. Both glomeromycotes and mucoromycotes might have been involved (Field et al. 2015c, d), or just glomeromycotes - mostly members of "basal" clades, far fewer Glomeraceae - may have been the early associates (Rimington et al. 2018: morphology of AM fungi various, but obviously there is little possibility of hyphae growing in intercellular spaces - see also F. A. Smith & Smith 1997 for variation in AM types), and similar associations were also found in hornworts. Interestingly, such associations tended to be specific in liverworts, but not in hornworts (Rimington et al. 2018). For more literature on the possibility that AM Glomeromycota were associated with the earliest land plants, see van der Heijden et al. (2015). Feijen et al. (2017) suggest that an asociation with Mucoromycotina might be the basic condition for land plants. This suggestion may depend both on the method of ancestral state reconstruction used (Bayesian) and patterns of relationships - here there was a tritomy made up of vascular plants, [mosses + liverworts], and hornworts in the latter two of which non-mycorrhizal groups were sister to clades in which no evidence was given that Mucoromycotina were basal (Feijen et al. 2017). However, given that patterns of association of fungi with extant bryophytes s.l. with fungi are complex and ideas of relationships within bryophytes and between bryophytes and other embryophytes are changing (see below); even if Mucoromycotina were the first fungal associates of liverworts, the establishment of plant-fungus relationships there may be independent of those in other land plants. Both Glomeromycota and Mucoromycotina may have been associated with the ca 407 Ma Horneophyton ligneri from the Rhynie Chert (Strullu-Derrien et al. 2014b), and such co-infections still occur (Field & Pressel 2018). ECM associations between liverworts and basidiomycetes (especially Serendipitaceae) and ascomycetes are likely to be secondary (Bidartondo & Duckett 2009; Weiß et al. 2016; Field et al. 2015d). There are further complications: Fungi can move to a liverwort, for example, from a tracheophyte (Ligrone et al. 2007) or in the opposite direction (Pressel et al. 2010: see also Bidartondo & Duckett 2009). For the (mostly ascomycetous, but some Sebacinales-Serendipitaceae) endophytic fungi also to be found in mosses and liverworts, see e.g. Stenroos et al. (2010), Pressel et al. (2010) and Weiß et al. (2016), and for general information, see elsewhere.

Lutzoni et al. (2018: bryophytes s.l. paraphyletic) suggest that the origin of glomeromycotes may have been (715-)659(-606) Ma, their diversification beginning (529-)484(-437) Ma, roughly contemporaneous with the early diversification of embryophytes (485-)482(-473) Ma; mucoromycotes, the sister group of glomeromycotes and possibly also involved in early AM symbioses, are ca 406 Ma, based on the divergence of Endogone. There are a number of other evolutionary events in fungi that may also be implicated in early embryophyte evolution, including specificially Tracheophyta, the stem age of which Lutzoni et al. (2018) estimate to be ca (466-)452(-438) Ma.

At least three genes involved in the establishment of mycorrhizae are likely to have been found in the common ancestor of land plants, and although the DM13 gene in particular seems to have other functions in many mosses, which are non-mycorrhizal, this gene from hornworts can rescue the effect of dmi3 mutants in Medicago truncatula (B. Wang et al. 2010, commentary by Bonfante & Selosse 2010). Strigolactones, involved in the establishment of AM associations, are known from some streptophytes (Charales) and seem initially to have been involved the control of rhizoid elongation (Delaux et al. 2012; Martin et al. 2017); a number of other genes involved in mycorrhizal establishment, the symbiosis signaling module, have a similar ancestry, although hardly surprisingly there have been elaborations in embryophytes (Delaux et al. 2015; see also Field & Pressel 2018). A karrikin receptor complex is involved in the initial interaction between plant and fungus (karrikin and strigolactone both have a butenolide element), and this complex is found widely in embryophytes (Delaux et al. 2012; Gutjahr et al. 2015).

Genes & Genomes. For the evolution of the land plant genome taking into account what is known of green algal streptophytes, see Rensing et al. (2007a) and especially Bowman et al. (2017). Renzaglia et al. (1995) discuss genome size. See Wicke et al. (2011) for nuclear ribosomal DNA organization and Leitch and Leitch (2013) and Szövényi (2016) for the small nuclear genomes of bryophytes s.l.. Overall, bryophytes s.l. have quite slow rates of speciation and of genome size evolution, as do pteridophytes and gymnosperms (particularly the latter), a siuation that does not really change until after the ANA grade of angiosperms (Puttick et al. 2015).

In Physcomitrella there are syntenic blocks in some chromosomes in which the genes are colinear across all embryophytes; the genes there are involved in land plant-specific cell growth and tissue organization (Lang et al. 2017). Pires and Dolan (2010) found that basic helix-loop-helix proteins, a class of transcription factors, diversified very early, while Volokita et al. (2010) discuss the GDSL-lipase gene family and its evolution. RNA editing, in which the organelle-targeted pentatricopeptide repeat proteins play an important role, is restricted to Embryopsida (Rüdinger et al. 2008). Sayou et al. (2014, see also the subsequent discussion) looked at the evolution of the LEAFY gene, usually in a single copy in embryophytes (but two in gymnosperms), yet variously involved in cell division and specification of floral identity. Its binding specificity is notably variable (promiscuous) in the hornworts, and although the evolutionary scenario suggested by Sayou et al. (2014) is based on the rather unlikely topology [hornworts [mosses [liverworts + vascular plants]]] (see also below), they suggest that whatever the topology, this promiscuity hypothesis is likely (see their Fig. S9). Sakakibara (2016) summarized studies suggesting substantial similarities in transcriptional regulation and cellular function in water-conducting cells in Physcomitrella and Arabidopsis thaliana, in regulatory mechanisms of stomata, and in rhizoid/root hair differentiation. Indeed, hairs on Coleochaete have similarities with the rhizoids of bryophytes (Graham et al. 2010a), while Mello et al. (2018) suggest possible connections between auxin response and roots or root-like structures across land plants in general. For more on the evolution of rhizoids and rhizoid-like structures, see Duckett et al. (2014), also below.

Much of interest is coming from the study of the evolution of individual metabolic pathways. Thus important signaling intermediates, the G-protein complex, is known from Chara and land plants (but not the green alga Micromeris), although elements of the pathway seem to be missing in some mosses and liverworts (Hackenberg et al. 2013). C3HDZ (class III homeodomain leucine zipper) genes, very important in development in euphyllophytes, etc., are also found in Chara (Prigge & Clark 2006; Floyd et al. 2006; Vasco et al. 2016). All land plants have some similar terpene synthase genes, but in addition there are microbial-type terpene synthase-like genes in all embryophytes except seed plants (at 3/7, hornworts have the lowest proportion of such genes), and these latter genes do not occur in streptophytes (Jia et al. 2016). Some of these microbial terpene synthase genes are similar to those of fungi and others to those of bacteria, and they were probably acquired by embryophytes through lateral transport (Jia et al. 2016, 2018), but other than the distinctive distribution of these genes, little is known about their evolution. Furthermore, triterpenoids are produced by CYP716 enzymes in all embryophytes except monocots (Miettinen et al. 2017). Various classes of phosphoprotein phosphatases also have interesting distributions, with the ApaH phosphatases currently being known from streptophytes only, while the ALPH class is absent from land plants (Uhrig et al. 2012). Several copies of LATERAL ORGANS BOUNDARIES DOMAIN genes are to be found in bryophytes, and they are also also known from Coleochaete and Spirogyra (Chanderbali et al. 2015). KNOX1 and KNOX2 genes occur in embryophytes, also in Klebsormidium, and KNOX1 regulates sporophytic meristems while KNOX2 suppresses the gametophytic developmental program in the sporophyte (Sakakibara et al. 2013, 2016). Two copies each of the plant homeobox KNOX genes, involved in meristem activity in vascular plants, and the MADS-box MIKC genes, i.a. floral identity genes, are found in mosses; subsequent duplication generated the diversity of these gene classes found in flowering plants in particular (Theißen et al. 2001).

Relatively little is known about plastid and other organelle inheritance in homosporous land plants in particular. Although placing "maternal inheritance" at this node may seem somewhat notional, organelle inheritance is usually uniparental (Wicke et al. 2011a; Barnard Kubow et al. 2016); it shifts at the gymnosperm node.

Details of gene and genome evolution in embryophyte plastids are given by Jansen et al. (2007), A. M. Magee et al. (2010), Wicke et al. (2011a), Civán et al. (2014) and Lemieux et al. (2016). Compared with their sister group, the [Zygnematales + Desmidales] clade, the chloroplast genome in basal land plants is relatively stable (e.g. Lemieux et al. 2016). For the evolution of polyplastidy, see de Vries and Gould (2017) - it seems to be associated with the loss/movement to the nucleus of minD, minE genes, which happened quite deep in the streptophyte clade. The chloroplast genes trnLUAA and trnFGAA are not associated in other green plants (Quandt et al. 2004). Ruhlman et al. (2015) outline the increase in complexity of the NADH dehydrogenase (NDH) complex in embryophytes; little of this been incorporated into the tree here. Details of the evolution of the chloroplast trnS and trnN genes in embryophytes are complex (Knie et al. 2014).

Most bryophytes s.l. have small mitochondrial genomes and little variation in gene order compared with mitochondria in vascular plants (Y. Liu et al. 2014a); for the evolution of the mitochondrial genome in basal land plants, see Knoop (2013). For intron distributions in mitochondrial genes, see Dombrovska and Qiu (1994), Qiu et al. (1998), and Regina et al. (2005). trnS(gcu) and trnN(guu) genes occur in the mitochondria of many streptophytes and also in liverworts; the former gene is also found in the mitochondria of the [monilophyte + lignophyte] clade, but not in Gnetales.

Chemistry, Morphology, etc.. For the hemicellulosic polysaccharide xyloglucans, see Del Bem and Vincentz (2010), Scheller and Ulvskov (2010) and Zabotina (2012) and especially Del-Bem (2018). Komatsu et al. (2014; see also Doi et al. 2006) discuss chloroplast movement, mediated by a single copy of the blue-light sensitive phototropin gene, in land plants.

See also Kenrick (2000) and Ligrone et al. (2012a), both morphology, O'Donoghue et al. (2013: xyloglucans), O'Rorke et al. (2015: sugars of pectic polysaccharides, interpretation of change depends very much on phylogeny), Taylor et al. (2009: fossils, inc. those of fungi associated with plants), Blackmore and Crane (1998: spore/pollen apertures), Brown and Lemmon (1990: callose and spore development, mono/polyplastidy, 2013 and references: sporogenesis), Brown et al. (2015: spore walls of basal members of the bryophyte clades), Wallace et al. (2011: spore and pollen wall development in land plants), Renzaglia and Garbary (2001: male gametes), Wastenays (2002: microtubules), Tomescu et al. (2014: apical cells of sporophyte), Waters (2003: molecular adaptation), Hedges et al. (2004: timing), Renzagia et al. (1995) and Hodges et al. (2012) cilia, Doyle (2013: reproductive features) and de Vries et al. (2016: plastid evolution).

Phylogeny. See Kenrick and Crane (1997), Nishiyama and Kato (1999) and Shaw and Renzaglia (2004) for early literature on bryophyte relationships. The three groups of bryophytes, mosses, liverworts and hornworts, are now nearly always found to be individually monophyletic, although this was sometimes not so for liverworts (Bopp & Capesius 1998: no vascular plants included; Quandt & Stech 2003, and references), and in some analyses Nickrent (2000) found that mosses were paraphyletic, although there was little support for this. There are seven main competing hypotheses of relationships between the three groups of bryophytes and vascular plants (e.g. see Nickrent et al. 2000; Puttick et al. 2018), and relationships here have been somewhat problematic (see also Cooper 2014), however, recent work by e.g. Wickett et al. (2014), Puttick et al. (2018) and Morris et al. (2018) in particular is clarifying relationships in this area.

1. Many studies support the set of relationships [liverworts [mosses [hornworts + vascular plants]]] (e.g. Kenrick & Crane 1997; Goffinet & Shaw 2009; Shaw et al. 2011 for literature; X.-X. Shen et al. 2017: evaluation of support for two hypotheses; Evkaikina et al. 2017); Qiu et al. (2006) confirmed these relationships using three different data sets. This may be the most supported hypothesis of relationships up to late 2017 (e.g. Lewis et al. 1997; Kelch et al. 2004; Wolf et al. 2006: many analyses, whole chloroplast genomes; Qiu et al. 2007; S. Li et al. 2013; Ruhfel et al. 2014: whole chloroplast genomes; Y. Liu et al. 2014b: mitochondrial nucleotide and amino acid data, other relationships very poorly supported; Magallón et al. 2013, 2016; Shimamura 2016 for literature). Dombrovska and Qiu (1994) had earlier outlined several lines of evidence such as the content of the inverted repeat and intron distributions that were consistent with the idea that liverworts were sister to all other land plants (see also Qiu et al. 1998b: mitochondrial introns; Antonov et al. 2000: cp rDNA ITS). This position is also favoured by an analysis of cpITS spacer sequences (Samigullin et al. 2002) and a complete plastome analysis (Karol et al. 2010). Kelch et al. (2004), using structural characters of the plastome, and Groth-Malonek et al. (2004, not all analyses; see also Knoop 2005), looking at trans-splicing mitochondrial introns, also suggested the same position for liverworts (see also Rydin & Källersjö 2002 and Karol et al. 2010, but neither in all analyses). Similarly, the distribution of group 2 nad4 mitochondrial introns suggests bryophyte paraphyly (Volkmar et al. 2012: Fig. 1 and references). Consistent with this hypothesis, the distribution of an extension of the chloroplast inverted repeat placed hornworts as sister to tracheophytes, as did the distribution of cell wall xylans (Carafa et al. 2005), and the distributions of some gene families also favoured this position (Szövényi 2016).

2. A [liverwort + moss] basal clade has sometimes been found (Ruhfel et al. 2014); see also Finet et al. (2010: hornworts not sampled), some analyses in Karol et al. (2010) and in B. Zhong et al. (2013b: Anthoceros sometimes sister to tracheophyta, sometimes sister to Lycopodiales), u.s.w.. In an extensive analysis of chloroplast genomes, Lemieux et al. (2016) recovered the relationships [[mosses + liverworts] [hornworts + the rest]], although support for the position of the sole hornwort included, Anthoceros, was weak. This topology has some support in Puttick et al. (2018). Group 1 introns in the mitochondrial cox1 and nad5 suggest a [mosses + liverworts] clade - or they were an embryophyte apomorphy that was later lost (Volkmar et al. 2012 and references).

3. Mitochondrial sequence data sometimes placed hornworts as sister to all other land plants (Sayou et al. 2014), and also in the analyses of Rai and Graham (2010), but there may be a rooting issue here. This position was also recovered by Hedderson et al. (1998: small subunit rRNA), with [mosses + liverworts] the next clade up, a topology that had some support in Puttick et al. (2018). Renzaglia and Garbary (2010) considered that the evidence for the same hornwort basal hypothesis was compelling, and they found this position to be quite strongly supported in their analysis of 123 morphological characters (Garbary & Renzaglia 1998), and several other studies have recovered this topology (Stech et al. 2003; Shanker et al. 2011 for references); relationships found by Nickrent et al. (2000) when 3rd codon position transitions were excluded were [hornworts [[mosses + liverworts] [vascular plants]]].

4. Mosses are sister to all other land plants in the analyses of Floyd et al. (2006, 2014) and Vasco et al. (2016), with either liverworts (Floyd et al. 2006) or hornworts (Vasco et al. 2016) sister to vascular plants, although basal land plant relationships were not the focus of these studies. There was no support for such topologies in Puttick et al. (2018).

5. In the chloroplast proteome analysis of Shanker et al. (2011: c.f. position of Huperzia) bryophytes were monophyletic. Nishiyama et al. (2004) also proposed that the three bryophyte groups formed a single clade; 51 genes from the entire chloroplast sequence were included, but taxon sampling was poor, e.g., no lycophytes were included. A similar grouping was also found in an analysis of the trnL intron (Quandt et al. 2004) and in another study that looked at many genes but with very skimpy sampling, that of Goremykin and Hellwig (2005). In the chloroplast genome Bayesian MCMC analysis of Civán et al. (2014) relationships were [[hornworts [mosses + liverworts]] [clubmosses, etc. + the rest]], although in a concensus tree [liverworts + hornworts] were sister to other embryophytes, within which relationships were pretty much scrambled. Cox et al. (2014) noticed that trees based on protein coding sequences and trees based on the proteins they coded differed in their topologies, and they suggested that there may have been synonymous substitutions in the sequences; they, too, argued strongly for the monophyly of bryophytes. A [liverwort + moss] clade was also recovered by Wickett et al. (2014) in transcriptome analyses, and in many analyses there, too, bryophytes (as [hornworts [mosses + liverworts]]) were monophyletic. In another comprehensive set of transcriptome analyses, this again was the predominant topology recovered (Puttick et al. 2018, see also Morris et al. 2018; Gitzendanner et al. 2018: chloroplast data; Cheng et al. 2018). This is the topology followed here. Puttick et al. (2018) found that hypotheses that did not have a [moss + liverwort] clade could be rejected, and in none of these last three studies did hypothesis 1 above have significant support.

Previous Relationships. In the old telling of the tale when there was a basal paraphyletic Bryophyta s.l., plants were lined up in a simple to complex sequence and evolution was read off the series accordingly. The emphasis was on the progressive acquisition of the features of vascular plants, e.g. of stomatal functions (Field et al. 2015b, but c.f. Renzaglia et al. 2017; Brodribb & McAdam 2017) and aspects of sporophyte development. Indeed, prior to iii.2018 on this site there was a group, the Stomatophytes, which included all embryophytes except the liverworts, that were characterized as follows:

Abscisic acid, L- and D-methionine distinguished metabolically; pro- and metaphase spindles acentric; class 1 KNOX genes expressed in sporangium alone; sporangium wall 4≤ cells across [≡ eusporangium], tapetum +, secreting sporopollenin, which obscures outer white-line centred lamellae, columella +, developing from endothecial cells; stomata +, on sporangium, anomocytic, cell lineage that produces them with symmetric divisions [perigenous]; underlying similarities in the development of conducting tissue and of rhizoids/root hairs; spores trilete; shoot meristem patterning gene families expressed; MIKC, MI*K*C* genes, post-transcriptional editing of chloroplast genes; gain of three group II mitochondrial introns, mitochondrial trnS(gcu) and trnN(guu) genes 0.

An [Anthocerophyta + Polysporangiophyta] clade was characterized as follows:

Gametophyte leafless; archegonia embedded/sunken [only neck protruding]; sporophyte long-lived, chlorophyllous; cell walls with xylans.

The task is now to see which of these features is best placed at the embryophyte node - "presence of stomata", for example, could be placed there, a feature like "loss of stomata" then being an apomorphy for the liverwort clade (Puttick et al. 2018).

I. BRYOPHYTES s.l.

Sporophyte-gametophyte junction with dead gametophytic cells, spaces containing mucilage; sporophyte of a single terminal sporangium; basal body with a proximal extension, basal body length; the angle of the spline with respect to the lamellar strip (= microtubule organizing centre, MTOC), etc. (for details, see e.g. Hodges et al. 2012, esp. Table 1)

Age. Around 506.5-460.5 Ma is the spread of ages for crown-group bryophytes s.l. suggested by Morris et al. (2018).

Evolution: Divergence & Distribution. When mosses were thought to be sister to vascular plants the conducting tissue in the centre of the stem in some moss gametophytes was homologized with the vascular tissue in the sporophytes of vascular plants (e.g. Mishler & Churchill 1984, 1985; Mishler et al. 1994). However, this was in part due to the way in which characters describing conductive tissue were conceptualized; there may be little reason to consider the conductive tissues of mosses and those of polysporangiophytes as having much similarity other than that due to their similar functions (e.g. Ligrone et al. 2000, 2002). Ligrone et al. (2002) found no great similarity between the water conducting cells of Takakia, the hydroids of other mosses, and the conducting tissues in Haplomitrium and metzgerialean liverworts (see also Doyle 2013).

However, recent work modifies such interpretations. Thus B. Xu et al. (2014) found that similar NAC transcription factor family genes were expressed during the development of the hydroids of mosses and the xylem of vascular plants, despite the difference in the generation in which they were expressed and in morphology (hydroids have neither pitting nor lignification), i.a. inducing cell death in both. Such NAC genes are uncommon in liverworts. Similarly, the formation of both sporophytic root hairs in Arabidopsis and gametophytic rhizoids and caulonemal cells of the protonema of Physcomitrella, all cells in which there is tip growth, involves the same regulatory gene network, perhaps independent recruitment and/or some kind of heterochrony/topy (e.g. Jones & Dolan 2012, p. 205: "transference of gene function"; Menand et al. 2007; Pires & Dolan 2010; Pires et al. 2013; see also Szövényi et al. 2010; Salazar-Henao et al. 2016: ROOT HAIR DEFECTIVE SIX-LIKE (RSL) genes involved; L. Huang et al. 2017). Menand et al. (2007), Jones and Dolan (2012), Kenrick and Strullu-Derrien (2014), Tam et al. (2015) and Hwang et al. (2017) also discuss the evolution of root hairs and rhizoids; the development of "tip-growing" cells like rhizoids, caulonema cells and root hairs is controlled by a similar auxin-regulated network - even fungal hyphae show similarities (Rounds & Bezanilla 2013, L. Huang et al. 2017 and references). In Physcomitrella caulonemal and rhizoid formation are inhibited by the same mutations (Rounds & Bezanilla 2013 and references). Sakakibara (2016) summarizes such findings; their extension to liverworts would be of considerable interest.

An argument can be made for the independent origins of PIN(= auxin efflux facilitators)-mediated auxin transport polarity in the sporophytes of mosses and vascular plants (Harrison 2017b: Fig. 1, Table 1), although there is also auxin transport in the gametophyte of bryophytes s.l. (apolar, plasmodesmatal) and polar transport of sorts in Klebsormidium and Chara, at least (references in de Vries & Archibald 2018). PIN plays a substantial role in the development of the Physcomitrella gametophyte, and its activity is evident even in apically-growing cells of the protonema (Bennett et al. 2014; Viaene et al. 2014: not in root hairs or pollen tubes in angiosperms). Interestingly, disruption of PIN in the sporophyte can lead to its branching (e.g. Bennett et al. 2014), and fossils like Partitatheca which lack vascular tissue but have stomata may represent a branched bryophyte sporophyte (see Harrison 2017 and literature).

If the three groups of bryophytes form a clade (see above), then interpreting the evolution of such features as stomata, cell division and the sporophyte shoot meristem becomes rather different, particularly if liverworts are sister to other bryophytes (e.g. Frank & Scanlon 2014; Lind et al. 2015), although as mentioned elsewhere, this seems rather unlikely. It may be that many of these features are apomorphies for extant land plants, being found either in the gametophyte (ancestral conducting tissue) or the sporangium (stomata), and the sporophytic generation became more elaborate in vascular plants, with rhizoids, stomata and vascular tissue forming a functional whole (Kenrick 2017). In the bryophytes s.l., however, features like "stomata in sporophytes" and "conductive tissues in gametophytes" may represent losses in liverworts and club mosses respectively. Optimization of such distinctive and important features as the presence of a sporangium with a seta and columella, stomata, trilete spores, polarity of transport of auxin in the sporophyte, etc., is not easy whether or not bryophytes are a paraphyletic group. Thus stomata are absent from liverworts and some of the basal clades of mosses, and their morphology and functioning can be quite distinctive (see stomatophytes); Merced & Renzaglia 2013; Haig 2013; Brodribb & McAdam 2017; Merced & Renzaglia 2018), perhaps suggesting their independent origin within mosses.

de Vries and Gould (2017) note variation in the number of plastids per cell in this clade. Qiu et al. (2006a, 2007 and references) suggested that a number of features of hornworts, particularly of the sporophyte, that may be synapomorphies of polysporangiates and hornworts. For the evolution of trilete spores, see e.g. J. A. Doyle (2012b) and Brown et al. (2015). I had "primary cell walls with xylans (xyloglucans with fucosylated subunits)" as an apomorphy of [Anthocerophyta + Polysporangiophyta] prior to July 2016, but xylans seem to be an apomorphy of land plants, although the nature of their side chains (charged vs uncharged) varies (Scheller & Ulvskov 2010; see also Sarkar et al. 2009; Zabotina 2012 and references). There is little borate cross-linked rhamnogalactan II in bryophytes (or in Chara), but there is much more in vascular plants where it cross-links primary cell wall pectins and seems to have been an important early step in the development of lignified cell walls (Matsunaga et al. 2004).

Thinking about diversification in bryophytes s.l. is difficult. Laenen et al. (2014) suggested that overall diversification rates in bryophytes s.l. (paraphyletic) were lower than in angiosperms in particular, but also lower than in ferns. This was despite species-level diversification rates of a number of extant groups being comparable to those in angiosperms, and they suggested that there had been "massive extinctions" (or numerous smaller extinctions?) in the past, which would explain the overall low level of bryophyte s.l. diversity (Laenen et al. 2014). However, some comparisons are tricky here. Laenen et al. (2014) found liverworts and mosses first diversified in the mid-Jurassic and mid-Cretaceous respectively, and rightly note that although Fiz-Palacios et al. (2011) found declining diversification rates for the former in the mid-Cretaceous, this may well have been because they used families as their unit of investigation, however, comparison of genus ages across bryophyta s.l. and angiosperms by Laenen et al. (2014) would seem to face similar problems.

Reproductive Biology. For a discussion about various aspects of sexual and asexual reproduction in the gametophytic generation of bryophytes, the latter very common and occuring in a variety of ways, see Maciel-Silva and Pôrto (2014). Since gametophytes spread vegetatively, it causes complications for sporophyte formation; if selfing takes place, the sporophyte is homozygous. Given this vegetative spread, dioecy is only a partial solution since sperm cannot swim very far, but dwarf males that are epiphytic on the females have evolved in a number of taxa (Haigh 2016).

Epiphytic bryophytes tend to be endosporic and have precocious germination, i.e. the spores are multinucleate, and sometimes they are green (Schuette & Renzaglia 2010; see also Nehira 1987).

Genes & Genomes. For post-transcriptional editing of the chloroplast genes, see Martín and Sabater (2010).

Knie et al. (2015) note that the cis- state of the nad2i542g2 intron of the mitochondrial genome is likely to be the ancestral condition for land plants. For the loss of the loss of the mitochondrial trnS(gcu) and trnN(guu) genes at this node, see Knie et al. (2014), and for the gain of three group II mitochondrial introns, see Qiu et al. (1998b).

Chemistry, Morphology, etc.. Poli et al. (2003), Sakakibara et al. (2008), Fujita et al. (2008) and Fujita and Hasebe (2009) discuss sporophyte growth and polar auxin transport in land plants, although sampling needs to be improved (what about liverworts?), and polar transport is not well developed in hornworts. Boot et al. (2012) have demonstrated polar auxin transport in the gametophyte of Chara.

For phytolith/silica deposition, see Thummel et al. (2019). Although quite widely spread, and indeed quite common in complex thalloid liverworts, Bryales and Polytrichales, phytolith/silica presence is unlikely to be the ancestral condition for any larger group in this area (Thummel et al. 2019).

Details of the sporophyte-gametophyte boundary are discussed by Duckett and Ligrone (2003). The sporangium develops from tiers of four cells (quadrants) which divide periclinally producing an outer amphithecium and an inner endothecium; the major groups of bryophytes differ in how the capsule wall and spores develop from these cells (Ligrone et al. 2012a for a summary). For spore morphology, see Qiu et al. (2012) and Blackmore et al. (2012) and references (correlation: monolete spores, successive sporogenesis, tetrads tetragonal/decussate, simultaneous sporogenesis).

Goffinet and Shaw (2009) and Shaw et al. (2011) provide much general information about the "bryophytes" as a whole.

ANTHOCEROPHYTA Stotler & Crandall-Stotler / HORNWORTS

Gametophyte: thalloid, (leafy), closely associated with N-fixing Nostoc, (glomeromycotes, mucoromycotes +); flavonoids 0; apical cell +, wedge-shaped, with four cutting faces; branching truly dichotomous; ventral mucilage clefts/cavities +, opening by stoma-like pore; vegetative cells monoplastidic, microtubules axial; antheridia in chambers [endogenous]; bicentriole pair formed in cell generation before spermatogenous cells; stellate array in basal body of cilium absent; male gametes bilaterally symmetrical, with a right-handed coil; archegonia embedded/sunken [only neck protruding]; sporophye-gametophyte junction with sporophytic haustorial cells [uniseriate/branched], transfer cells in gametophyte; sporophyte: long-lived, chlorophyllous, with plasmodesmatal auxin transport, lacking apical cell, foot ovoid-bulbous; stomata +, 51-81 μm long; basal meristem active for an extended period, seta short; sporangium dehiscing by 2 longitudinal slits; amphithecium producing archesporial tissue; spores maturing from base of capsule to top, elaters +, spirally thickened, multicellular; xylans in walls of spores and elaters; n = 5, genome [1C] 0.085-0.28 pg/(156-)244(-714) Mb; mitochondrial rpl2 gene 0.

11/300. World-wide.

Age. The age of this clade has been estimated at around 290 Ma (Villarreal & Renner 2012), or ca 170 Ma, the overall variation being (245.2-)228.9, 160.2(-107.1) Ma (Villarreal et al. 2015a).

LEIOSPOROCEROTOPSIDA Stotler & Crandall-Stotler

Gametophyte: Nostoc in branching schizogenous strands in the centre of the thallus, mucilage clefts only in young uninfected plants; antheridia up to 70/chamber; Sporophyte: spore tetrads bilateral alterno-opposite, spores "minute", monolete, surface smooth.

1/1: Leiosporoceros dussii.

ANTHOCEROTOPSIDA Jancz.

Neochrome +; chloroplast grana lack highly curved end membranes, channel thylakoids connect adjacent granal stacks; gametophyte: plastids (>1/cell, mitosis still monoplastidic), pyrenoids +/0; Nostoc in spherical colonies, mucilage clefts persist; antheridia 1-6/chamber; sporophyte: (first division of zygote vertical - Anthoceros); (stomata 0 if sporophytes more or less enclosed); 1-many chloroplasts/cell; white line-centred lamellae 0?; spores ornamented, (multicellular, chlorophyllous); phototropins lack introns; elevated rate of RNA editing.

10/215: Phaeoceros (), Dendroceros (62). World-wide.

Age. The age of this clade may be (399-)306, 248(-173) Ma (Villarreal & Renner 2012) or 261-186 Ma (Laenen et al. 2014).

Evolution: Divergence and Distribution. Most diversification within Anthocerotopsida has taken place within the last 100 Ma or so, or even within the Caenozoic (Villarreal & Renner 2012; Villarreal et al. 2015a).

Ecology & Physiology. Neochromes, a chimaeric photoreceptor in which red-sensing phytochrome and blue-sensing phototropin are fused into a single molecule, have been found in all hornworts sampled (Leiosporoceros not studied: F.-W. Li et al. 2014).

Dendroceros is the only dessication-tolerant hornwort, it is also epiphytic and has green, multicellular spores, features perhaps associated with the epiphytic habitat (Schuette & Renzaglia 2010).

For details of stomatal opening, see Renzaglia et al. (2017; also Merced & Renzaglia 2018; Pressel et al. 2018). As Renzaglia et al. (2017) note, the stomata open once, remain open, and appear to assist in the drying of the capsule contents (the sporogenous tissue is initially bathed in mucilage, and intercellular spaces in the sporophyte are initially filled with liquid, not gas. Consistent with this function, neither stomatal size or number respond to variations in CO2 concentration (Field et al. 2015b). Some taxa lack stomata on their sporangia, but in that case the sporangia are more or less enclosed by the gametophytic involucre (Renzaglia et al. 2017).

Pyrenoids (for which, see Hanson et al. 2014) have evolved five times or more in the hornworts between 101 and 18 Ma (and also subsequently been lost several times); they vary considerably in morphology. Their repeated evolution in Anthocerotopsida may be an example of a "tendency", however, what is causing their gains and losses is unclear (Villarreal & Renner 2012). Pyrenoids may increase the concentration of CO2 around the chloroplasts and so increase the efficiency of photosynthesis (E. Smith & Griffiths 1996).

For the association of hornworts with the nitrogen-fixing Nostoc, see Rai et al. (2000), and for what is known about Nostoc and nitrogen fixation, see Santi et al. (2013).

Plant-Animal Interactions. Labandeira and Currano (2013) note a variety of herbivory types on Middle Devonian liverworts.

Bacterial/Fungal Associations. Endogone-like fungi (Mucoromycotina) are associated with some hornworts (Bidartondo et al. 2011; Rimington et al. 2014), and are quite common there, as are Glomeromycota (Pressel et al. 2010). The two may form mycorrhizal associations with the same species, but whether either or both mycorrhizal association was ancestral in the group is unclear (Desirò et al. 2013); glomeromycote associations, at least, seem rather casual (Pressel et al. 2010; Rimington et al. 2018).

In most taxa Nostoc enters the gametophyte through mucilaginous clefts (Adams 2002; Adams & Duggan 2008).

Genes & Genomes. The extensive RNA editing in Anthoceros and its relatives is at codon positions that are otherwise universally conserved in land plants (Duff et al. 2007). There seems to have been between 1-6 duplications in each of the three groups of GDSL lipases somewhere around here (Volokita et al. 2010). Hornworts may have sex chromosomes, and so, as might be expected, polyploidy is uncommon here (Bowman et al. 2017 and references).

Anthoceros has lost the chloroplast rps15 gene (Martín & Sabater 2010) and the IR has expanded somewhat (Villarreal et al. 2013).

There has been extensive loss of protein-coding genes in the mitochondrion, i.e. ca 1/2, versus less than 1/3 in other land plants (Y. Liu et al. 2014b). Leiosporoceros has an intron in the mitochondrial nad5 gene, as do Anthoceros and immediate relatives (Villarreal et al. 2013).

Chemistry, Morphology, etc.. The cell walls of spores and their elaters contain xylans, an important polysaccharide in secondary cells walls of vascular plants, but unknown in other bryophytes (Carafa et al. 2005). The elaters are multicellular structures of tapetal origin, quite unlike those of Equisetum (Pacini & Franchi 1991).

The pores of the mucilage clefts of hornwort gametophytes are probably not homologous with stomata (Adams 2002; Adams & Duggan 2008).

See also Ligrone et al. (2000) for general information, Villarreal et al. (2010) for a summary of our understanding of hornworts, Vaughn et al. (1992) for the distinctive and variable chloroplasts, Brown and Lemmon (2013) and Schuette and Renzaglia (2010: Dendroceras) for sporogenesis, and for antheridia, see Duff et al. (2004) and Cargill et al. (2005).

Phylogeny. Relationships within hornworts are still unclear in part. Leiosporoceros may be sister to all other hornworts, although the extensive RNA editing in other members of the clade obscures this position in some analyses (e.g. Duff et al. 2007); it has many distinctive features (see above). See also Stech et al. (2003), Duff et al. (2004) and Villarreal and Renner (2012) for relationships. Anthoceros and its possible segregate Folioceros are sister to remaining hornworts, and they have black or dark spores, etc. (in this they are like Leiosporoceros). Villarreal and Renner (2014) discussed the limits of and relationships around Nothoceros.

Classification. Several classifications of this clade have appeared recently (Frey & Stech 2005a; Stotler & Crandall-Stotler 2005; Duff et al. 2007); they tend to be rather elaborate. Söderström et al. (2016) provide a classification down to the level of species.

[MARCHANTIOPHYTA + BRYOPHYTA s.str.] / SETAPHYTA

Gametophyte: perichaetial leaves +; sporophyte: caspule with stalk.

MARCHANTIOPHYTA / LIVERWORTS

Gametophyte: thallus simple; lunularic acid +; rhizoids smooth, living/(pegged, dead); perforate water conducting cells +; membrane-surrounded oil bodies + (0); microtubule organizing centres + [MTOCs polar organizers]; cell walls with relatively little cellulose; sporophyte: plasmodesmatal auxin transport apolar; foot ± conical to spheroidal, cell divisions uniform, sporangium initially enclosed, seta [stalk] evanescent, forming by cell elongation after the sporangium develops; sporangium columella 0, wall 1(2-4) cell layers across, opening by four slits; endothecial cells producing archesporial tissue; meiosis usu. polyplastidic; elaters +, unicellular; mitosis with polar MTOCs; spore walls with more or less continuous parallel lamellae at maturity [?level]; nuclear genome [1C] (0.21-)1.2-9.44(-20.46) pg/(206-)1,844(-20,010) Mb.

Ca 4,000-7,486 spp. (see Söderström et al. 2016).

Age. The crown group is dated to (509-)484(-452)Ma by Cooper et al. (2012) and 509-486Ma by Laenen et al. (2014), although Heinrichs et al. (2007) suggested an age of (410.5-)407.6(-404.7) Ma, very much younger.

From fossil evidence, liverworts or liverwort-like plants were present by the Ordovician, and liverworts may have been common through the Silurian and Devonian (Graham et al. 2011); see also Heinrichs et al. (2007) for an evaluation of the identity of fossils purported to be liverworts.

1. Haplomitriopsida Stotler & Crandall-Stotler

Gametophyte axial; fungal symbiont mucormycotina; thallus apical cell tetrahedral, (stem anatomy complex - Treubia), mucilage copious [stalked slime papillae]; stem erect, leaves spiral, (unlobed), costa 0/± prostrate, leaves two-ranked; (rhizoids 0 - Hapolomitrium); plant dioecious; bracts surrounding sporophyte 0; meiosis monoplastidic [?all]; (spores in diads); embryo haustorium?; distinctive blepharoplast; nuclear genome [1C] 4.35-10(-10.05) pg/4254-9794 Mbp.

3/16. More or less world-wide, scattered.

Age. Crown-group Haplomitriopsida are dated to (396-)353(-316) Ma (Newton et al. 2007) or (469-)414(-352) Ma (Cooper et al. 2012).

[Marchantiopsida + Jungermanniopsida]: gametophyte dorsiventral; leaves 0; late development of placental wall ingrowths; (meiosis monoplastidic), (spore development endosporic).

Age. This node is dated to before the Middle Devonian (475-)442(-408) Ma (Cooper et al. 2012), while Heinrichs et al. (2007) suggest an age of (382.8-)372.6(-362.4) Ma (similar ages also in Newton et al. 2007) and B. Zhong et al. (2014b) an age of (569.7-)375.8(-172.5) Ma, which pretty much covers the waterfront.

2. Marchantiopsida Cronquist

Thallus branching truly dichotomous, ventral scales in two rows; gemmae in receptacles [gemma cups]; paraphyses +; sporocytes often lacking lobing, monoplastidic, (MTOCs at nuclear envelope); placental transfer cells variable; no RNA editing in organellar genomes.

Ca 340 spp.

Age. Crown-group Marchantiopsida are estimated to be (322-)284(-251) Ma (Cooper et al. 2012), (268-)248(-231) Ma (Newton et al. 2007), 269-207 Ma (Laenen et al. 2014) or (365-)295(-250) Ma (Villareal A. et al. 2015b).

There are fossils of complex thalloid liverworts from the Triassic and mid-Cretaceous (references in Villareal A. et al. 2015b).

2a. Blasiidae He-Nygrén

Fungal symbiont 0, association with Nostoc; thallus simple, margins lobed; plant dioecious; capsules ellipsoid, walls multistratose; meiosis monoplastidic; nuclear genome [1C] 0.5 pg/488 Mbp.

2/2: North Temperate.

2b. Marchantiidae Engler

Thallus complex, with air chambers, air pores to the outside; transfer cells in both sporophyte and gametophyte; apical cells of sporophyte 0; sporocyte meiosis monoplastidic.

Age. The age of this clade is (328-)263(-226) Ma (Villareal A. et al. 2015b).

2b1. Neohodgsoniales D. G. Long

Fungal symbiont mucoromycotes and glomeromycotes; thallus with compound air pores; rhizoids dimorphic, some dead; archegoniophore/carpocephalum +, branched; ?nuclear genome.

1/1: Neohodgsonia mirabilis: New Zealand and Tristan da Cunha.

2b2. Sphaerocarpales Caveo

Fungal symbiont 0; (thallus lacking air chambers/pores), (oil bodies 0), pegged rhizoids 0, ?= smooth, (carpocephala +); sporangium wall breaks down at maturity, elaters 0; ?nuclear genome.

4/20-30. Europe, North America, Australia, South Africa (!once found - Monocarpus).

2b3. The Rest: ?

(Plant annual); fungal symbiont glomeromycote (and mucoromycote), thallus differentiated (not), photosynthetic filaments + (0); air pores simple or compound; rhizoids dimorphic, some pegged, dead; archegoniophore/carpocephalum +, simple, with furrows in which there are pegged rhizoids.

3. Jungermanniopsida

(Plant epiphytic); fungal symbiont mucoromycotina, dikaryan, glomeromycote, or 0; fructan sugars accumulated; thallus +, simple/leaves +, (2-)3-ranked, lobed, costa 0; ?rhizoids; placental fungal haustoria +, transfer cells only in sporophyte (0).

Ca >4,000 spp.

Age. Differences in suggested ages for crown-group Jungermanniopsida are rather great, almost 100 My: e.g. (425-)390(-353) Ma (Cooper et al. 2012), (335-)328.5(-322) Ma (Heinrichs et al. 2007) or (331-)292(-262) Ma (Newton et al. 2007).

Pelliidae He-Nygrén et al.

Plant a simple thallus; nuclear genome [1C] 0.55-9.44(-20.46 - Phyllothallia) pg/894-9237(-20006) Mbp.

Metzgeriidae Bartholomew-Began

Plant a simple thallus; nuclear genome [1C] 0.74-9.24 pg/724-9037 Mbp.

Jungermanniidae Engler

Plant radial, leafy, cutting faces of apical cell at 120o, leaves succubous (incubous/transverse/plant a simple thallus); perigynium around sporangium.

Evolution: Divergence and Distribution. Prototaxites, a trunk-like structure up to 8.8 m long and 1.37 m diameter, although often much smaller, apparently is the largest land organism of the Late Silurian to Late Devonian 420-370 Ma. Its identity has been the subject of much discussion, and on balance, the evidence suggests that it is neither a fungus nor a lichen, but is made up largely of rhizoids of marchantioid liverworts and of remains of their fungal associates, all rolled together (Graham et al. 2010a).

Villarreal A. et al. (2015b: q.v. for dates) discuss evolution in the complex-thalloid clade, i.e. Marchantiopsoda; overall molecular evolution (plastid and mitochondria) was slow compared to that of other hepatics, except in Cyathodium, an annual (a reversal), but morphological evolution was less so. Given the topology of the tree (e.g. the position of Neohodgsonia is important), where features of the carpocephalum (the archegonial receptacle after fertlization) and those involved in the differentiation of the thallus - the variation is extensive - are to be paced on the tree is unclear, but the pattern of gains and losses will be complex (Villareal A. et al. 2015b).

Heinrichs et al. (2007) discussed the evolution of the ca 4,500 species of leafy liverworts, suggesting possible divergence times for the clades (see also Newton et al. 2007; Coooper et al. 2012), and Wilson et al. (2007a, b) discuss the diversification of Lejeuneaceae in particular. Much of the diversification of Porellales and Jungermanniales, both leafy liverworts epiphytic on bark and leaves of flowering plants, has occurred since the evolution of angiosperms (Ahonen et al. 2003; Forrest & Crandall-Stotler 2004), although the clades involved are (much) older (Cooper et al. 2012). Indeed, although many liverwort families had diverged by the end of the Cretaceous, they have diversified considerably since (Cooper et al. 2012). Note that although ages in Newton et al. (2007), Heinrichs et al. (2007) and Cooper et al. (2012) can differ greatly, the overall conclusions of the authors differ less; additional dates may be found in these publications.

The L and D isomers of methionine are treated identically metabolically in liverworts, almost alone in land plants (Kenrick & Crane 1997).

Ecology & Physiology. Understanding the ecophysiology of the first liverworts is a challenge. Field et al. (2012, 2014 and references) discuss details of the nature of the mutualistic association between fungus and plant in some liverworts, including Haplomitriopsida (see also Ligrone et al. 2007; esp. Field et al. 2015c). The association of Mucoromycotina with Haplomitrium benefits both partners, and since this liverwort lacks rhizoids the fungus is particularly important in nutrient uptake (Field et al. 2014). See elsewhere on this page for ideas of the identity or the original fungal associates of liverworts and their possible environmentl effects. Such fungi are involved in soil formation in liverwort communities, along with cyanobacteria, etc., and although soil accumulation in thinner than in moss communities, at least in Iceland (ca 1.5 vs 3.5 cm), clay minerals like smectite, also accumulate. (Mitchell et al. 2016)

The echlorophyllous Cryptothallus [= Aneura] mirabilis is the only mycoheterotrophic liverwort, indeed, it is the only mycoheterotrophic bryophyte s.l.. It can grow up to 20 cm below the surface of Sphagnum-dominated peat bogs and is associated with hyphae of the ectomycorrhizal basidiomycete Tulasnella. The latter is simultaneously associated with Betula or Pinus from which the liverwort indirectly obtains its carbon (Wickett et al. 2008 and references; Merckx et al. 2013a).

Extant Marchantia, at least, is mixotrophic (Hata et al. 2000; Graham et al. 2010a, b).

Quite a number of liverworts, including epiphytes, tolerate extreme dessication, although most lack internal water-conducting cells (Ligrone et al. 2000). Dessication tolerance can be constitutive or induced (Oliver et al. 20105; Gaff & Oliver 2013).

Reproductive Biology. The dead, pegged rhizoids to be found in marchantialean liverworts with complex thalli may help ensure the water supply to the stalked carpocephala (Duckett et al. 2014).

Plant-Animal Interactions. For a summary of herbivory and galling, including examples from the Middle Devonian where some cells of the fossils perhaps contained oil and represent defence against herbivory, see Labandeira et al. (2013). Caterpillars of Micropterigidae, a basal, jawed, lepidopteran clade, are common on Conocephalum conicum in Japan, and this group also includes species eating other liverworts elsewhere, also angiosperms, detritus, etc. (Imada et al. 2011; Ragier et al. 2015 and references).

Bacterial/Fungal Associations. All the major groups of fungi that form mycorrhizal associations with land plants are associated with liverworts (Read et al. 2000; Duckett et al. 2006b; Pressel et al. 2010; Bidartondo et al. 2011; Field et al. 2014, 2015c). Some liverworts in the basal pectinations are associated with both Glomeromycota and Mucormycotina (Kottke & Nebel 2005; Field et al. 2014; Rimington et al. 2014), but this association may subsequently have been lost (Duckett et al. 2006b); the associations of Jungermanniopsida-Jungermanniales with ascomycetes are more than 250 Ma (Pressel et al. 2008b). The fungus may move from these liverworts to seed plants (Pressel et al. 2010: see also Bidartondo & Duckett 2009). Thus the ascomycete Rhizoscyphus [= Hymenoscyphus] ericae is very commonly an associate of the hair roots of North Temperate Ericaceae, and it also forms mycorrhizal associations with Jungermanniales-Schistochilaceae and other leafy liverworts; colonization of the liverwort is by their rhizoids (Duckett & Read 1995; Upson et al. 2007; Pressel et al. 2008b). Of course, these indirect associations with seed plants have little necessarily to do with early liverwort-fungus associations (e.g. Duckett & Read 1995). Members of a clade of Sebacinales-Serendipitaceae are associated with liverworts (Weiß et al. 2016), and this is also likely to be a relatively recent connection. Colonization of the plant is of the Paris type, with extensive intracellular hyphal growth producing coils, but sometimes neither arbuscules nor vesicles (Field et al. 2015d).

Epiphytic Porellales lack fungus associations, and in general epiphytic or epilithic liverworts are often not associated with fungi (Pressel et al. 2010), a tendency also evident in angiosperms. Similarly, liverworts in more or less transiently wet and nutrient-rich habitats also lack mycorrhizae, as do ferns like Equisetum and Salviniales (Pressel et al. 2016) and aquatic angiosperms.

A number of endophytic fungi grow in Marchantia polymorpha, a species that rarely has mycorrhizae. Nelson et al. (2018) detail the variety of positive and negative effects these fungi have on the plant, effects that partly depend on the identity of other fungi in the plant, and partly even on the age of the association - and the effects may also be quite different on other plants with which the fungi are associated.

The morphology of the gametophyte of Treubia depends on whether or not it has associated fungi (Field et al. 2015a; Rimington et al. 2017).

The mycoheterotrophic Cryptothallus [= Aneura] mirabilis is associated with the ectomycorrhizal basidiomycete Tulasnella (Cantherellales) (Kottke & Nebel 2005; Wickett & Goffinet 2008; Wickett et al. 2008; Imhof et al. 2013; for Tulasnella and liverworts, see Oberwinkler et al. 2017). Aneura is also associated with basidiomycetes (Pressel et al. 2010).

Blasia (Marchantiopsida) fixes nitrogen by virtue of of its association with Nostoc (Rai et al. 2000; Warshan et al. 2018 for the relationships of N-fixing Nostoc), but it does not form any associations with fungi (Field et al. 2015c).

Genes & Genomes. Endopolyploidy has rarely been detected in liverwort nuclei (Bainard & Newmaster 2010a, b), although the nuclei of the smooth rhizoids of at least some liverworts are highly endopolyploid (Duckett et al. 2014). Variation in genome size, even at the infraspecific level, is quite considerable, but there seems to be no correlation with chromosome numbers (Bainard et al. 2013). There is also little evidence of genome duplications in the Marchantia genome, and Bowman et al. (2017) suggest that this is because of the early evolution of sex chromosomes in the group; presence of sex chromosomes is associated with genome stability (see also hornworts). If polyploidy does occur, the species usually becomes monoecious (Bowman et al. 2017 and references). For the Marchantia polymorpha genome, see Bowman et al. (2017). There is substantial variation in the size of the nuclear genome, that of Marchantiopsida being particularly small, but few species have been examined (Villareal A. et al. 2015b).

Marchantiopsida have lost to ability to carry out RNA editing of the organellar genes (Rüdinger et al. 2008).

Plastid genomes in liverworts generally seem to be conserved and are rather similar to those in their streptophyte ancestors, if often somewhat smaller, while their GC contents are rather variable, that of Haplomitrium and Aneura being notably high (Y. Yu et al. 2019b). For the chloroplast genome of the mycoheterotroph Aneura mirabilis, see Wickett et al. (2008). The genome is more or less normal in size, although it has functionally lost around 25 genes, much in line with gene losses in parasitic angiosperms (see also Bellot & Renner 2015).

Chemistry, Morphology, etc.. Details of meiosis, whether there is one or more chloroplasts, etc., vary considerably in liverworts, and some derived taxa are quite like vascular plants in these respects (Brown & Lemmon 2008, 2013). There is monoplastidic meiosis in Monoclea, Haplomitrium and Blasia (Renzaglia et al. 1994a; Brown & Lemmon 2011a), but there is also polyplastidic meiosis as well as intermediate forms. Buschmann et al. (2016) discuss microtubles and cell division in Marchantia.

For features of Haplomitriopsida, see Duckett et al. (2006a), for those of Marchantia, see Shimamura (2017: also much else), for Marchantiidae, see Flores et al. (2018), for apical cell division, see Piatkowski et al. (2013), for the development of the unicellular elaters, see Renzaglia et al. (1997) and Crandall-Stotler and Stotler (2000), for spore walls, see Wellman et al. (2003), and for rhizoids, see Duckett et al. (2014).

Phylogeny. Morphological studies indicated that Sphaerocarpos might be sister to all other liverworts (Crandall-Stotler & Stotler 2000). However, molecular data suggest rather different relationships, and that genus is now included in Marchantiopsida (see also Forrest & Crandall-Stotler 2004, 2005; He-Nygrén et al. 2004; Qiu et al. 2006); the long branches associated with Haplomitrium and Treubia cause their position in the tree to be somewhat migratory. He-Nygrén et al. (2006: 3 chloroplast and 1 nuclear genes, morphology) outline the phylogeny and classification of liverworts, finding a basic structure [Treubiopsida [Marchantiopsida + Jungermanniopsida]]. This basic topology is confirmed by Forrest et al. (2006: five genes, all three compartments, good sampling, esp. of thalloid liverworts, 2015), Volkmar and Knoop (2010) and Cooper et al. (2012). N.B.: In older literature, Treubiopsida = Haplomitriopsida.

Marchantiopsida include Blasia, Sphaerocarpos, etc., although support for the inclusion of the former in this clade could be improved (but c.f. some analyses in Forrest & Crandall-Stotler 2004, esp. 2005; Qiu et al. 2007: Blasia sister to other Marchantiopsida; see also He-Nygrén et al. 2004; Volkmar et al. 2011 ). Cooper et al. (2012; see also Forrest et al. 2015) also found Blasia - with Cavicularia - to be sister to the other Marchantiopsida.

Within Jungermanniopsida, simple-thallus groups are paraphyletic with respect to the speciose and monophyletic leafy liverworts, within which Pleurozia is sister to the rest or near-basal (e.g. He-Nygrén et al. 2004; Davis 2004: P. grouped with some simple-thalloid genera; Masuzaki et al. 2010; Volkmar et al. 2011; Cooper et al. 2012: extensive study). These general relationships were also recovered by Qiu et al. (2007). Leafy liverworts have been studied by Cooper et al. (2011: the speciose Lepidoziaceae), and Gradstein et al. (2003) and Yu et al. (2013), both Lejeunaceae.

Classification. See Frey and Stech (2005b) and Crandall-Stotler et al. (2009) for phylogeny-based classifications and also Söderström et al. (2016) for a classification (followed above) that goes down to species.

BRYOPHYTA s.s. / MOSSES.

Gametophyte: several developing from a single spore; leafy, radial, axial, cutting faces of apical cell at 136o [most]; cells ± differentiated, leaves +, unistratose, costa +; perichaetium +; sporophyte: early embryo spindle-shaped, foot ± elongate-tapering-pointed, seta developing from basal meristem [between epibasal and hypobasal cells], indurated, tissues differentiated; PIN[auxin efflux facilitators]-mediated polar auxin transport, calyptra +, persistent; endothecium also producing archesporial tissue; placenta with transfer cells in sporophyte alone; MTOC from plastids, becoming diffuse, perinuclear; perine + [?level]; endopolyploidy widespread; x = 7, nuclear genome [1C] (0.17-)0.38-0.92(-2.05) pg/(170-)504(-2,004) Mb. 31 orders.

Age. Crown-group mosses may be (400-)379(-362) mMa (Newton et al. 2009).

The separation of Takakia and Sphagnum is dated to 319-129 Ma (Shaw et al. 2010a).

1. Takakiopsida Stech & W. Frey / Takakiophytina

Gametophyte: initially ?thalloid, mycorrhizal; oil bodies + [?membrane-surrounded], rhizoids 0, vegetative cells monoplastidic [mitosis, too, monoplastidic]; cutting faces of apical cell at 120o; leaves forked, ± 3-ranked; perforate water conducting cells +; plant acrocarpous; perichaetial leaves 0; sporophtyte: capsule dehiscence spiral; stomata 0; sporocytes unlobed, spores not trilete; n = 4.

1/2. East Asia, west North America.

[Sphagnopsida [Andreaeopsida + The Rest]]: gametophyte: initially a filamentous persistent protonema; rhizoids branched, multicellular [septate, branches very fine]; mycorrhizae 0; leaves spiral, unlobed; cp ITS3 differences.

Age. The crown-group age of this clade is (360-)352(-344) Ma (Y. Liu et al. 2014a) or 448.5-344.5 Ma (Morris et al. 2018: Sphag. + rest), but see below for basal relationships.

2. Sphagnopsida Ochyra / Sphagnophytina

Gametophyte: fructan sugars accumulated; protonema thalloid, rhizoids +, in mature gametophyte 0; branches in axillary fascicles, spreading and pendant; leaf cells dimorphic [groups of empty and hyaline cells surrounded by strands of chloroplast-containing cells]; sporophyte: sessile, borne on gametophytic pseudopodium; foot from hypobasal cell, bulbous, placental transfer tissue 0; stomata + ["pseudostomata"], stomium 0; capsule dehiscence subapical and transverse ["operculate"], explosive; columella massive, overarched by spores; placental transfer tissue 0; amphithecium alone producing archesporial tissue; sporocytes unlobed, spore wall multilayered; n = 38, 42.

5/250. World-wide.

Age. Crown-group Sphagnopsida are around 104-30 Ma (Shaw et al. 2010a) or a mere ca 25 or even 14 Ma (Shaw & Devos 2014).

455-454 Ma fossils of leaves with a distinctive cell structure remarkably similar to that of extant Sphagnum are known from the Ordovician deposits in Wisconsin (Cardona-Correa et al. 2016). Ivanov et al. (2018) discuss the detailed morphology of Palaeozoic protosphagnalean mosses, and although there are some similarities with the cell arrangement of Sphagum, the similarities between the two have different origins, cell expansion versus cell division.

[Andreaeopsida + The Rest]: ?

3. Andreaeopsida J. H. Schnaffner / Andraeophytina

Gametophyte: initially thalloid; plant acrocarpous; leaf costa +/0; sporophyte: capsule sessile, borne on gametophytic pseudopodium; stomata 0; columella overarched by spores; (calyptra 0); dehiscence down four (eight) vertical slits; spores not trilete, exine initiated as globules, white-line centred lamellae 0, germination endosporic.

2/110: Andreaea (110). World-wide, rather scattered, esp. cool southern/circum-Antarctic.

4. Bryophytina

Gametophyte: (L and D isomers of methionine identical metabolically - Mnium); plant acrocarpous (pleurocarpous); (rhizoids 0 - Haplomitrium); hydroids + [cells dead, no contents], leptoids containing refractive spherules; paraphyses + [multicellular chlorenchymatous hairs mixed with gametangia] (0 - Buxbaumia); sporophyte: (first division of zygote longitudinal - Funaria); stomata + [Oedidpodium up] (0), (guard cell single, binucleate); capsule dehiscence transverse, peristome +; spores hilate; PEP α subunit rpoA gene 0 (+).

Age. The crown-group age of this clade is perhaps 602-488 Ma (Laenen et al. 2014: none of the three clades above was included).

[Oedopodipsida [Tetraphidopsida + Polytrichopsia]]: nematodontous

BRYIDAE Engler / Bryopsida

Peristome double, teeth alternating [arthrodontous].

BRYANAE

HYPNANAE

plant pleurocarpous [female gametangia on short lateral ± leafless branches].

[Hypnodendrales [Ptychomniales [Hypopteygiales [Hookeriales + Hypnales]

[Ptychomniales [Hypopteygiales [Hookeriales + Hypnales]

[Hypopteygiales [Hookeriales + Hypnales]

[Hookeriales + Hypnales]

HOOKERIALES

HYPNALES

42 families, 430 genera, ca 4,000 species.

Evolution: Divergence & Distribution. Mosses in China showed weaker latitudinal diversity gradients than did liverworts, perhaps because the former can handle a wider diversity of environments than the latter, a number of which are epiphyllous and need humid conditions; moss diversity was linked in part to local habitat heterogeneity (S.-B. Chen et al. 2015).

Dicranidae/haplolepidious taxa are a diverse but species-poor group compared to the ca 12,000 species of Hypnanae/hypnalian pleurocarpous/arthrodontous mosses (Cox et al. 2010); Newton et al. (2007, 2009; see also B. Zhong et al. 2014b) give dates for many clades. Coudert et al. (2017) looked at the evolution of branching patterns in the gametophytes in a study that focussed on Bryidae and Hypnanae in particular. i.e. on pleurocarpous mosses, although a number of acrocarps, including a few Dicranidae, were also examined.

Within the speciose pleurocarpous mosses - about 40% of all mosses - diversification seems to have been early and rapid, clades diverging in the early Cretaceous, but since then there has been semi-stasis (Kürschner & Parolly 1999; Shaw et al. 2003b; Newton et al. 2006, 2007), although there may also have been more recent ([post-]Cretaceous) diversification as well. The diversification of Hypnales, with ca 1/3 of all moss species, may be associated with a genome duplication - or several small duplications (M. G. Johnson et al. 2016).

There is strong geographical signal in the phylogeny of Polytrichopsida, clades being largely south or north temperate (Bell & Hyvönen 2010: intergeneric hybridization?). The closest relative of the South American endemic dung moss Tetraplodon fuegianus is a population from West North America (Washington), divergence occurring ca 8.6 Ma, interestingly, overall geographical relationships in Tetraplodon were [Papua* [Nepal [Alaska [[Washington + South America*] [other N. Hemisphere samples]]]]], asterisks representing clades other than T. mnioides (Lewis et al. 2017).

Since some combination of Andreaea, Takakia and Sphagnum and immediate relatives are at the base of the moss phylogenetic tree and all have distinctive morphologies, apomorphies for mosses as a whole have been unclear, however, Y. Liu et al. (2019) found quite strong support for the topology followed here; characters are often mentioned. Huttunen et al. (2013: focus on Plagiothecaceae) optimize the evolution of a number of features in both Hypnales and Hookeriales. For distinctive features of Sphagnales, see Shaw et al. (2016b).

Ecology & Physiology. Dessication tolerance is common in mosses, even in taxa like Sphagnum, however, neither Sphagnum nor Takakia shows the extreme dessication tolerance that is quite common in other mosses, where it can be constitutive or inducable (Oliver et al. 2005; Gaff & Oliver 2013); see also papers in Plant Ecol. 151(1). 2000. The antheridia remain functional if they dry slowly and then are subsequently rehydrated (Stark et al. 2016 and references). Interestingly, Slate et al. (2019) suggest that dessication-rehydration cycles may have an appreciable effect on soil nutrient balances since both carbon and nitrogen are lost from rehydrating mosses in appreciable quantities in the water used for rehydration (the equivalent of rainfall) so boosting their amounts in the soil; mosses that were continuously hydrated did not show this effect.

The absence of arbuscular mycorrhizae in mosses has occasioned comment, and it has been suggested that their finely branched rhizoids are a substitute (Field et al. 2015d). Such fungi are involved in soil formation in liverwort communities, but obviously not in mosses, which develop a thicker layer of soil (mineral) grains trapped by the plants, but less in the way of clay minerals, etc. (Mitchell et al. 2016).

Mosses are important components of tundra and boreal forest biomes (for the bryosphere, see Lindo & Gonzalez 2010) and can make up a substantial proportion of the biomass in boreal forests (Wardle et al. 2013). Hummock-forming Sphagnum in particular is often dominant in communities from poor fens to the forest floor in tundra and boreal biomes. Indeed, rich fens in such areas commonly progress from so-called brown mosses like Drepanocladus and Calliergon and a pH of ca 6 to Sphagnum-dominated bogs with a pH of 4 or a little more - oligotrophification and acidification is mediated by Sphagnum, a process that proceeds fastest in cool, wet conditions (Kuhry et al. 1993).

The ecophysiology of Sphagnum has been studied in some detail. Particularly in exposed areas, damage from high-light conditions becomes important, as with other mosses, and drying out may also occur - but water reduces CO2 diffusion by a factor of 104, so potentially affecting plant productivity. However, there are methanotropic bacteria that live in the hyaline cells in the leaf; these bacteria prefer damper conditions, oxidizing methane from decomposing peat and producing CO2 which is then utilized by the plant. This can provide up to a third of the CO2 supply for the moss, and all 23 of the species of Sphagnum at a site in Finland could convert methane into a source of CO2, so this capability may be ubiquitous in the genus, while in Scorpidium scorpioides, which grows in similar habitats, the figure is up to 70% (Larmola et al. 2010; Hájek 2014). For the microbiome of Sphagnum, integral to the functioning of the organism and its ecosystem, see Bragina et al. (2014) and Graham et al. (2017). Sphagnum can also take up mono- and disaccharides - the monosaccharide glucose is preferred - and so is mixotrophic, considerably enhancing the biomass of the plant and perhaps particularly important at times of carbon and/or light limitations, but at least some other mosses and liverworts can also do this (Graham et al. 2010b).

Different species of Sphagnum tend to grow in the hummocks and hollows of peat bogs, and these preferences correlate with different growth rates, rates of peat breakdown (Turetsky et al. 2008), phylogeny, and so on. There is extensive pH variation in these bogs, and different species prefer different pHs, but here there is no large-scale correlation with phylogeny (M. G. Johnson et al. 2014). Sphagnum-dominated fens in northern Alberta may not be very productive in terms of gross primary productivity, however, the plants start photosynthesizing early in the year, etc., so overall net carbon productivity may be higher than in, for example, Carex-dominated rich fens (Flanagan 2014; see also Ragoebarsing et al. 2005). Breakdown of moss peat in general, and Sphagnum peat in particular, tends to be slow, and differences in the rate of peat breakdown can be linked to variation in carbohydrate metabolism within the genus (Turetsky et al. 2008). Cell wall pectin-like polysaccharides and glycuronoglycones (their dispersed form = sphagnan) have antimicrobial and tanning activity (see the Tollund Man!), while in their cell-wall bound form they acquire nutrients efficiently because of their high cation exchange capacity. Thus in peat they sequester nitrogen, for example, making it unavailable to microorganisms, so contributing to the resistance to decay of Spahgnum litter which has a low respiration rate - microorganisms grow there poorly (Painter 1991; Hájek et al. 2011). From this point of view, sphagnans take over the role of lignin; removing lignins affects the rate of Sphagnum breakdown very little, but it considerably increases it in Leucobryum and Polytrichum, mosses that behave more conventionally (Hájek et al. 2011).

The discovery of Sphagnum-like fossils in Ordovician rocks ca 455 Ma suggests that the ecological equivalents of modern Sphagnum peatlands may have been around for a long time, and this has considerable implications for carbon fixation and sequestration in general, the evolution of Sphagnum-associated biota in particular, etc. (Graham et al. 2013, 2017, esp. Cardona-Correa et al. 2016). Carpenter et al. (2015) found spores of Stereisporites - linked to Sphagnum - to be very common in fire-prone heathlands in Central Australia 75-65.5 Ma, while Daly et al. (2011) suggest that Sphagnum-type mosses were components of the peats produced by mire vegetation in northern Alaska ca 60 Ma that later were converted to coal. However, the beginning of the diversification that gave rise to extant species of peat-forming Sphagnum has been dated to as late as the mid-Miocene ca 14 Ma (Shaw et al. 2010a; Shaw & Devos 2014), while the species that are major accumulators of peat today make up only a clade, albeit a major clade, within Sphagnum, species of subgenus Rigida and the segregate genera growing on moist to wet rock or soil (Shaw et al. 2016a, b). Indeed, the subpolar-temperate peatlands where Sphagnum is now so common seem to have developed within the last 17,000 years, largely as temperatures warmed after the last glacial maximum or, in the case of peats in the West Siberian lowlands, the triggering factor seems to have been increased precipitation (Morris et al. 2018). The balance between peat formation and sequestration is complex. As Treat et al. (2019) note, over the last 130,000 years northern peatlands in general have higher peat production in interglacials, however, peatlands near the coast may be buried as sea levels rise (and so the carbon is sequestered), but on the other hand during colder periods mineral deposition, as by glaciers, may result in the burial and sequestration of recently formed peat, but the fall in sea levels may expose new areas suitable for peat formation. Finally, a genome duplication in Sphagnum dated to (221-)197(-173) Ma may have been involved is its rise to dominance in peatlands (Devos et al. 2016) - and this gets us back to the dating of Sphagnum diversification, the fossil record of the genus, and where it might have grown when the earth was much warmer than it is now. Altogether a bit dizzying.

Spores of Spagnum ca 680 years old are able to germinate (Bu et al. 2017). Gametophytes of Chorisodontium aciphyllum, a moss from Antarctic islands, grew successfully after being frozen in moss banks for ca 1,600 years (Roads et al 2014), and even older ages seem quite likely.

In northern ecosystems a few species of feather mosses, pleurocarpous Hypnales like Hylocomium splendens and Pleurozium schreberi, form close associations with the cyanobacterium Nostoc, nitrogen moving from the latter into the former (Bay et al. 2013); sufficient phosphorus and in particular molybdenum is essential (Rousk et al. 2016). However, the further movement of nitrogen in the ecosystem is rather unclear (Rousk et al. 2013, 2016; Lindo et al. 2013).

Mosses are known for their tolerance of sometimes quite extreme dessication, despite their lack of an epidermis and much in the way of a cuticle, and also (depending on the species) to high NaCl, bisulphite (SO2), etc., concentrations (Rensing et al. 2007a and references; Hoang et al. 2009).

Gametophyte vascularization is particularly well developed in Polytrichopsida, and the leptoids apparently transport organic molecules (Ligrone et al. 2000).

Reproductive Biology. For the recurrent evolution of dioecy in mosses - at least 133 times - see McDaniel et al. (2013); dioecy may be accompanied by sexual dimorphism of the plants. Reversal to hermaphroditism is less common, but diversification may be higher in hermaphroditic clades (McDaniel et al. 2013).

When the antheridium dehisces, the contents are released as a single sperm mass, and subsequently the individual sperm disperse (Stark et al. 2016 and references). Very small arthropods are attracted to volatile compounds produced by Ceratodon purpureus and are involved in the transfer of sperm to the archegonia (Rosenstiel et al. 2012; see also Cronberg et al. 2006).

For a comprehensive study of the evolution of capsule shape in mosses, see Rose et al. (2016). At least sometimes shape changes are associated with changed speciation rates, as is the adoption of the pleurocarpous habit.

The stomata of Sphagnum lack both pores and airspaces immediately under the guard cells. The stomata open as they lose turgor and help the capsule to dry out; as the capsule buckles the stomium pops off and the spores are shot into the air (Duckett et al. 2009; see also Merced & Renzaglia 2018).

Bacterial/Fungal Associations. Bay et al. (2013) discuss associations between Nostoc and some pleurocarpous mosses in boreal forests; other blue-green algae are also involved (Rousk et al. 2013 for a review). Warshan et al. (2018) discuss the relationships of N-fixing members of Nostoc; the ability to form such associations may be older than the age of Bryophyta.

Functional mycorrhizal associations, i.e. associations that are involved in the exchange of nutrients, are very rare/absent in mosses, and most moss-fungus associations involve parasitic fungi (Read et al. 2000; Davey & Currah 2006). Fot the possibility that Buxbaumia is mycoheterotrophic, see Imhof et al. (2013). Other endophytic fungi may also affect moss growth and ecology (Read et al. 2000; Davey & Currah 2006).

Genes & Genomes. Genomes in mosses are small, 1C values being less than 1.4 pg (Bennett & Leitch 2005); I do not know what the sizes are in liverworts, etc.. The rate of molecular evolution is slow in the three genes from all three genomic compartments examined (Stenøien 2008: liverworts, hornworts, and lycophytes not examined). Lang et al. (2018) noted that the chromosomes of Physcomitrella were rather differently organized than those of seed plants, for instance, eu- and heterochromatin are fairly evenly distributed on the chromosomes, unlike the common condition in flowering plants, the centromeres were not so obvious as in vascular plants, although Copia-type transposable elements were concentrated there, however, the chromosomes appeared to be monocentric.

The base chromosome number for mosses may be x = 7 (Rensing et al. 2012). Endopolyploidy is widespread in mosses, although not in Sphagnum (Bainard & Newmaster 2010a, b). Nevertheless, genome duplications are reported from this genus (Devos et al. 2016), and overall, genome duplications were detected in 3/5 of the mosses examined by Lang et al. (2018). Duplications might have been involved in the rise of Sphagnum to dominance in peatlands, and the genes retained after this event are similar to those retained after similar events in other plants, i.e. those involved in ion channels or signal transduction pathways. Physcomitrella differs somewhat in this respect (Szövényi et al. 2015; Devos et al. 2016, and references), but there is evidence of two fairly recent duplications there (Rensing et al. 2007b; Lang et al. 2018). Yue et al. (2012) found that a number of genes in Physcomitrella patens had come from fungi and bacteria by lateral transfer; how widely they might be distributed in other mosses is unknown.

There is a very large (ca 71 kb) inversion of the chloroplast genome in Funariaceae (which includes Physcomitrella), Disceliaceae, and Encalyptaceae, all Funariidae, although Gigaspermaceae lack this inversion (Goffinet et al. 2007). For the loss of the PEP α subunit-encoding rpoA gene in many mosses, see Goffinet et al. (2005); it is absent in Tetraphis and Diphyscium, present in Buxbaumia and Polytrichum.

For variation in the mitochondrial nad2 and -5 genes, see Beckert et al. (1999, 2001).

Chemistry & Morphology. As moss rhizoids branch, they become progressively narrower, about as wide as a fungal hypha; for the distinctive nature of their cytoplasm, see Pressel et al. (2008a) and Field et al. (2015d). In the stomata of Funaria, at least, the stomium is encircled by a single cell that has two nuclei (Sack & Paolillo 1983; Merced & Renzaglia 2016: spacing), however, in other mosses the stomata have two guard cells (Merced & Renzaglia 2013, 2018). For spore wall morphology of Sphagnum, see Brown et al. (1982); some layers, e.g. the translucent layer, may be unique.

For a general entry into the literature, see Goffinet et al. (2004) and for information about pleurocarp mosses, see Newton and Tangney (2007); for sporogenesis, see Brown and Lemmon (1984: Andreaea, 2013), for apical cell division, see Piatkowski et al. (2013), for sieve elements/leptoids, see Scheirer (1990), and for placental tissue, see Carapa et al. (2003).

Phylogeny. Sphagnum, Andreaea and Takakia, the latter initially thought to be a liverwort (see Renzaglia et al. 1997), are all at the base of the moss tree, Sphagnum and Takakia perhaps being sister taxa and Andreaea sister to remaining mosses (e.g. Cox et al. 2004; Qiu et al. 2006, 2007: rather strong support; Volkmar & Knoop 2010; Shaw et al. 2010b [perhaps]; S. Li et al. 2013: 9 loci; Rose et al. 2016; Evkaikina et al. 2017: Andreaea not sampled). However, Takakia has a region in the cpITS3 sequence that is very like that of all other land plants but is deleted in other mosses, from this evidence alone, Takakia would be sister to all other mosses (Samigullin et al. 2002), and Y. Liu et al. (2019) found the same topology and with quite strong support. Moreover, most of the rest of the tree Liu et al. (2019) produced, based on the analysis of a large number of genes from all three compartments in 142 species from all bar one of the moss orders, had quite good support. Beckert et al. (1999: Takakia not included) found Polytrichum interpolated between Sphagnum and Andreaea. Recent work suggests relationships may be [Takakia [Sphagnum [[Andreaea + Andreaeobryum] [Oedopodium [[Polytrichopsida + Tetraphidopsida] Bryopsida]]]]] (Chang & Graham 2009, esp. 2011, 2014; Rose et al. 2016: but c.f. in part above); a [Takakia + Sphagnum] clade was recovered only in some reconstructions. The odd, almost leafless Buxbaumia may be sister to all other Bryopsida.

See Shaw et al. (2010b; also Shaw et al. 2003a for morphology, 2010a, 2016a) for relationships in the Sphagnum et al. clade. The very distinctive Sphagnum leucobryoides (= Ambuchanania) was described only some twenty five years ago (Yamaguchi et al. 1990), and it and a very few other species of Sphagnum (placed in genera like Flatbergia) are outside Sphagnum s. str., which makes up the bulk of the clade.

Within the remaining mosses, Chang and Graham (2009, esp. 2011) and S. Li et al. (2013) found Oedopodium to be sister to the others. See Cox et al. (2010) for a phylogenetic study of mosses focussing on genera and families. Using variation in mitochondrial genes, Beckert et al. (2001) found Buxbaumiales to be paraphyletic immediately above the basal grade of mosses and below Bryopsida. Wahrmund et al. (2010) used a new mitochondrial locus to investigate relationships; the position of Timmia was particularly unclear.

For relationships in Dicranidae (haplolepidious mosses), see Stech et al. (2012). Bell et al. (2007) discuss the phylogeny of the early diverging pleurocarp clades, and adjust their taxonomy accordingly, while Buck et al. (2005) discuss the phylogeny of Hookeriales. Most branch lengths in the speciose Hypnales are short (Huttunen et al. 2012).

Classification. For a classification of mosses based on phylogeny, see Shaw and Goffinet (2000, also Goffinet & Buck 2004), and for an (over)classification of Sphagnum s.l., see Shaw et al. (2010b).

POLYSPORANGIOPHYTA

Sporophyte well developed, chlorophyllous, branched, branching dichotomous, potentially indeterminate; hydroids +; stomata on stem; sporangia several, terminal; spore walls not multilamellate [?here].

Evolution: Divergence & Distribution. Note that the number of spores produced per sporangium is roughly the same here as in bryophytes s.l. (Qiu et al. 2012).

Bacterial/Fungal Associations. Vesicular arbuscular mycorrhizae have been found in the axes of Aglaophyton major from the early Devonian ca 400 Ma (Remy et al. 1994). A variety of fungal associations have been found in fossils of early polysporangiophytes, whether in gametophyte or sporophyte, and sometimes the fungi pervade the whole plant (e.g. Strullu-Derrien et al. 2014b). Because true roots had not yet evolved, they are best called paramycorrhizal associations (Kenrick & Strullu-Derrien 2014; for early fungal associations, see also Rimington et al. 2014).

Classification. Polysporangiophyta technically lack vascular tissue of any sort, and they include tracheophytes as a subgroup; for the two broken out, and protracheophytes, with hydroid and leptoid conducting cells, paratracheophytes and eutracheophytes also separated, see Gerienne et al. (2016).

II. TRACHEOPHYTA / VASCULAR PLANTS

Sporophyte long lived, cells polyplastidic, photosynthetic red light response, stomata open in response to blue light; plant homoiohydrous [water content of protoplasm relatively stable]; control of leaf hydration passive; plant endohydrous [physiologically important free water inside plant]; PIN[auxin efflux facilitators]-mediated polar auxin transport; (condensed or nonhydrolyzable tannins/proanthocyanidins +); borate cross-linked rhamnogalactan II, xyloglucans with side chains uncharged [?level], in secondary walls of vascular and mechanical tissue; lignins +; roots +, often ≤1 mm across, root hairs and root cap +; stem apex multicellular [several apical initials, no tunica], with cytohistochemical zonation, plasmodesmata formation based on cell lineage; vascular development acropetal, tracheids +, in both protoxylem and metaxylem, G- and S-types; sieve cells + [nucleus degenerating]; endodermis +; stomata numerous, involved in gas exchange; leaves +, vascularized, spirally arranged, blades with mean venation density ca 1.8 mm/mm2 [to 5 mm/mm2], all epidermal cells with chloroplasts; sporangia in strobili, sporangia adaxial, columella 0; tapetum glandular; sporophyte-gametophyte junction lacking dead gametophytic cells, mucilage, ?position of transfer cells; MTOCs not associated with plastids, basal body 350-550 nm long, stellate array in transition region initially joining microtubule triplets; archegonia embedded/sunken [only neck protruding]; embryo suspensor +, shoot apex developing away from micropyle/archegonial neck [from hypobasal cell, endoscopic], root lateral with respect to the longitudinal axis of the embryo [plant homorhizic].

Note: In all node characterizations, boldface denotes a possible apomorphy, (....) denotes a feature the exact status of which in the clade is uncertain, [....] includes explanatory material; other text lists features found pretty much throughout the clade. Note that the precise node to which many characters, particularly the more cryptic ones, should be assigned is unclear. This is partly because homoplasy is very common, in addition, basic information for all too many characters is very incomplete, frequently coming from taxa well embedded in the clade of interest and so making the position of any putative apomorphy uncertain. Then there are the not-so-trivial issues of how character states are delimited and ancestral states are reconstructed (see above).

Age. Clarke et al. (2011: many other estimates) suggested an age for this clade of (456-)446(-425) Ma, similar to the estimate of 451-431 Ma of Morris et al. (2018: stem of 25-60 Ma, other dates also suggested). Somewhat younger ages for vascular plants, (434.3-)424-421.6(-416.2) Ma, can be found in Magallón et al. (2013), while Larsén and Rydin (2015) suggest ages of ca 432 Ma, largely in line with fossil-based ages (Kenrick et al. 2012) which were used in the calibrations (a similar age in Villarreal & Renner 2014); ca 419 Ma is the estimate in Evkaikina et al. (2017). However, P. Soltis et al. (2002: variety of estimates) suggested an older crown age of (813-)603(-393) Ma. Silvestro et al. (2015) estimated that vascular plants were (449-)433.5(-424) Ma, and other estimates are broadly similar, e.g. (463.5-)429(-400) Ma (B. Zhong et al. 2013b), 458-442 Ma (Barba-Montoya et al. 2018), (449-)435(-426) Ma (Lutzoni et al. 2018: note topology, "differentiation" of tracheophytes (467-452(-438) Ma) and (440-)431.5(-426) Ma (Testo et al. 2018b).

Evolution: Divergence & Distribution. Pryer et al. (2004b) provide a useful summary of the evolution of vascular plants; see also e.g. Kenrick and Crane (1997), also Doyle (2013), etc..

Rothfels et al. (2015a) suggested that lycopods, ferns and other plants with abiotically mediated fertilization evolved reproductive incompatability only slowly, hence they have an intrinsically slow rate of speciation. On the other hand, there is no evidence that the small spores of homosporous lycopods and ferns are particularly associated with long distance dispersal, Gondwanan vicariance patterns being evident in distributions. At the small scale allopatric speciation is likely and outcrossing is the norm, indeed, speciation mechanisms/patterns here and in flowering plants seem to be similar, and speciation in the Andes, at least, can be quite rapid (Testo et al. 2018b, 2019 and literature).

The basic pattern of early development of the sporophyte, exoscopic or endoscopic, is outlined by Niklas and Kutschera (2009). Their scenario involves both gains and losses of endoscopic development; the optimisation here involves parallel gains, but is equally parsimonious. For sporopollenin in land plants, see above. Baucher et al. (2007) discuss vascular evolution, Johnson and Renzaglia (2009) that of the embryo, while Arens et al. (1998: Virtual paleobotany laboratory) is a valuable web resource.

Ambrose and Vasco (2015) suggested that the apical meristems of shoots of vascular plants would best be thought of as multicellular structures with cytohistochemical zonation (for plasmodesmata, see Imaichi & Hiratsuka 2007, c.f. interpretation of apical meristems there; Evkaikina et al. 2017). However, there is discussion as to the occurrence of apical cells in polysporangiophytes, and it is common to distinguish between tracheophytes with apical meristems of a single cell and those with meristems of several cells (Kato & Akiyama 2005; Imaichi 2008; Evkaikina et al. 2017: monoplex vs simplex/duplex). Interestingly, the expression patterns in the single apical cells of Selaginella and Equisetum are quite different (Frank et al. 2015). This may suggest that the apical cells in ferns s.l. and Selaginella evolved independently - or that the differences between the two reflect the long period they have been separated, some 400 Ma or more (see above for ages; Frank et al. 2015; Evkaikina et al. 2017). Details of the construction of the apical meristem varies within lycophytes and they correlate with the richness of plasmodesmatal connections between the cells: many plasmodesmata - meristem a single cell, few plasmodesmata - meristem a group of cells (Imaichi & Hiratsuka 2007; see also discussion in Evkaikina et al. 2017: Fig. 1). Fern apical meristems for the most part are multicellular, the apical initial(s) only rarely dividing (Ambrose & Vasco 2015; Vasco et al. 2016; c.f. Evkaikina et al. 2017: unicellular for the most part). D'Amato and Avanzi (1968; see also Gifford 1985) noted that the apical cells of Equisetum early became polyploid and did not subsequently divide (c.f. in part Gifford et al. 1979). In fern gametophytes the apical cell is functional for a short while only, and then the apical region converts to a multicellular meristem (Takahashi et al. 2009, 2015 and references). Details of the evolution of the organization of the stem apex will in part depend on whether bryophytes s.l. are monophyletic or paraphyletic; Evkaikina et al. (2017) suggest that having a single, apical cell is plesiomorphic for land plants. For further discussion, see elsewhere.

Roots, and so root caps, etc., are thought to have evolved two or more times in vascular plants (e.g. Raven & Edwards 2001; Jones & Dolan 2012; Kenrick & Strullu-Derrien 2014; Hetherington et al. 2016b: Fujinami et al. 2017; Hetherington & Dolan 2018), although details are still not well understood. Kenrick and Strullu-Derrien (2014) distinguished between rhizoid-based rooting systems (RBRS) in which branching of the axes was exogenous and dichotomous/bifurcating, and roots proper, in which branching was endogenous. Fujinami et al. (2017) described four distinct variants in the apical meristems of extant lycophyte roots, including one with an apical meristem (Selaginella) approaching the common type of organization in monilophytes, and another with a quiescent centre (= a common initial zone) or something similar approaching one kind of root apex organization found in seed plants - the last is basically an open meristem, the other three described are variants of closed meristems, and some kind of closed meristem could be the ancestral condition for extant vascular plants. For convenience I have put the acquisition of roots with a root cap and root hairs at this node; note that seed plant roots/root caps sense and respond to gravity in a different way (via statocysts, far faster) that do the roots of other vascular plants (Y. Zhang et al. 2019). Little is said about the presence of cuticle on roots (e.g. Kenrick 2013; Kenrick & Strullu-Derrien 2014), and although Matsunaga and Tomescu (2016) were unable to detect a cuticle in fossils of early vascular plant that they examined, they thought that it could have been there.

Within angiosperms, gene families involved in root development are conserved, the great majority being found in all six angiosperms studied, and this is true of genes involved in root development in Arabidopsis in particular (L. Huang & Schiefelbein 2015). Remarkably, around 82% of the gene families that are expressed in all angiosperm roots are also expressed in roots of Selaginella, even root cap-associated genes. Although details of the functional similarities of these genes is unknown, there may have been parallel evolution of root-associated genes in lycophytes and other vascular plants, or the dissimilarity of roots of lycophytes and other vascular plants was, at one time at least, not that great, some sort of roots occuring in the ancestor of extant vascular plants (L. Huang & Schiefelbein 2015). Similarities in the acropetal flow of auxin in the roots of lycophytes and seed plants are probably the result of independent evolution (Sanders et al. 2011). Furthermore, at least some details of the molecular development of root hairs and rhizoids are similar across all vascular plants, with ROOT HAIR SPECIFIC genes interacting with others to stimulate root hair development (Hwang et al. 2017; see also L. Huang et al. 2017), and some similarities extend still more broadly into the bryophytes (see above). Root hairs may initially have beeen able to differentiate from almost any cell, i.e., there were no trichoblasts, visibly distinguishable precursors of root hairs (Clowes 2000), however, whatever the root hair distribution pattern, there is specificity of gene expression because of a conserved transcription factor and cis-element (D. W. Kim et al. 2006).

Sporophyte root morphology and mycorrhizal relationships are connected. Pressel et al. (2016; see also Rimington et al. 2017) suggest that fungal associations have been progressively lost in monilophytes, perhaps because the epiphytic habit is common in Polypodiales (mycorrhizae tend to be less common in such habitats) and they have thin, wiry roots ≤1 mm across in which mycorrhizal associations are uncommon, unlike the thicker, more fleshy roots that may harbour mycorrhizae found in some other monilophytes. However, where the feature "wiry roots" is to be placed on the tree is unclear. I have provisionally put it at the level of vascular plants as a whole, since thin roots are common in Selaginellaceae, Lycopodiaceae and Equisetaceae, fleshy roots/rhizomes are found in some basal fern clades, but thin roots are found in most of the rest (Pressel et al. 2016). Any glomeromycotes associated with basal clades of vascular plants seem to be Glomeraceae, a derived group, compared with the associations with more basal glomeromycotes that occur in liverworts and hornworts (Rimington et al. 2018).

Floyd and Bowman (2006), Boyce (2008a), Boyce and Leslie (2012) and others emphasize the diversity of leaf morphologies, growth forms, etc., to be found in non-angiospermous plants in general. However, all leaves, whether megaphylls or microphylls, can be thought of as structures with "a determinate growth programme [added] to the indeterminate apical growth programme" (Harrison et al. 2005b: p. 509). Similar but independently recruited developmental mechanisms may be involved in the evolution of microphylls and megaphylls (Harrison et al. 2005b), as in many other aspects of evolution in land plants (Pires & Dolan 2012).

Zimmermann's (1930, see also 1952, 1965) telome theory suggests that euphylls are the result of overtopping, planation and webbing of a stem/branch system, and this theory has been influential, if in detail it is no longer tenable (Kenrick 2002; Beerling & Fleming 2007). The development of euphylls or megaphylls, often quite large structures the vascular supply of which is associated with gaps in the central stele, may be quite different from that of microphylls, vascularized structures that are not associated with leaf gaps in the central stele (e.g. Bower 1935; Floyd & Bowman 2006). Thus the leaves of lycophytes are characterized by having an intercalary meristem and being supplied by a single vein that does not leave a gap in the central stele when it departs (see Kaplan 1997, vol. 2: chap. 16, vol. 3: chap. 19, 2001 for leaf morphology), while megaphylls are determinate organs with ad/abaxial identities, the vascular bundles supplying them leaving "gaps" in the central stele when they depart (see e.g. Sporne 1965, but c.f. Kaplan 1997, vol. 3: chap. 19, 2001 in particular, see also Harrison et al. 2005; Boyce 2005a: summary of earlier literature, 2008; Tomescu 2008, 2009; Sanders et al. 2007, 2009; Corvez et al. 2012). However, details of the evolution of the euphylls that are supposed to characterise the Euphyllophyta/[Monilophyta + Lignophyta] clade are unclear. H. Schneider et al. (2009: p. 461 and references) suggest that mega/euphylls arose once, and can be characterized by apical/marginal growth, apical origin of the venation, determinate growth, etc., while Floyd and Bowman (2007) suggested that they evolved independently in seed plants and monilophytes - an estimated 3-6 times in the latter alone - and perhaps elsewhere as well, and D.-M. Wang et al. (2015) proposed that there had been four independent origins of megaphylls in ferns/monilophytes, progymnosperms, and seed plants by the end of the Devonian (see also Kenrick & Crane 1997; Boyce & Knoll 2002; Osborne et al. 2004; Gensel & Kenrick 2007; Tomescu 2009; Niklas & Kutschera 2009; Sanders et al. 2009; Galtier 2010; Corvez et al. 2012: at least two origins; Tomescu et al. 2014).

Indeed, Harrison et al. (2005b), Floyd et al. (2014), Vasco et al. (2016) and others are helping to develop an understanding the developmental background of euphylls or megaphylls or whatever they are called, i.e., of leaves in general. Some of the genes involved in megaphyll formation may have evolved long before megaphylls appeared, but with different functions then, (e.g. Beerling 2005a, b; Floyd 2006; Vasco et al. 2016 and references). Floyd et al. (2014) place the origin of the post-translational negative regulators of apical meristem and leaf development, the LITTLE ZIPPER (ZPR) proteins, at the [Monilophyta + Lignophyta] node; they are part of the C3HDZ (class III homeodomain leucine zipper) stable of genes, which may initially have been expressed in sporangia in embryophytes from bryophytes on up (Vasco et al. 2016 and references). After gene duplication of the C3HDZ gene in the [Monilophyta + Lignophyta] clade, one of the copies degenerated and became the ZPR gene, and it and C3HDZ genes together were involved in the evolution of euphylls. Floyd et al. (2014) could not find ZPR genes in lycophytes.

Interestingly, there may be more commonality between microphylls and megaphylls than might appear, and Vasco et al. (2016) suggest that C3HDZ genes may have been coopted in both megaphyll and microphyll development. Following Zimmermann, both megaphylls and microphylls can be thought of as modified sporangia. In lycophytes one of a pair of sporangia became a microphyll (see also Crane & Kenrick 1997) with C3HDZ genes being expressed in their development, while in the [Monilophyta + Lignophyta] clade more complex telome systems make up the leaf and C3HDZ genes are initially expressed throughout the young primordium, but later on only on the adaxial surface and then they are involved in the development of dorsiventrality (a neofunctionalization - Vasco et al. 2016; see also Scarpella & Meijer 2004 for the radial/dorsiventral sequence). Overall, there is a very complex pattern of gains, modifications, and losses of genes (Vasco et al. 2016; see also Evkaikina et al. 2017).

A number of features characterizing various levels of the tracheophyte clade have evolved more than once, heterospory, secondary thickening and the presence of an initial free nuclear/coenocytic phase in the development of the megagametophyte being examples. Heterospory has evolved several times in land plants (e.g. Bateman & DiMichele 1994; Qiu et al. 2012). Seed plants are heterosporous, and spore-bearing and photosynthetic leaves seem to be quite different, while in other heterosporous vascular plants there is no fundamental dissimilarity between the two (e.g. Kaplan 1997, vol. 3: chap. 19; Boyce 2005b). Haig and Westoby (1988, 1989) outline the conditions conducive for the evolution of heterospory; it may first have appeared in Devonian vegetation which was becoming denser and casting more shade - it helped in the establishment of the germinating plants (Petersen & Burd 2018 and references). In ferns, at least, Duckett and Pang (1984) suggested that temporal dioecy in the gametophyte, e.g. archegonia being produced first, antheridia later, might be a prelude to heterospory. The pattern of secondary growth varies considerably. The vascular cambium is usually unifacial, producing xylem internally only, and there are usually no anticlinal divisions of the cambial cells. Sphenophyllales alone outside the seed plant lineage may have developed a bifacial vascular cambium (e.g. Rothwell et al. 2008b; Spicer & Groover 2010; Hoffman & Tomescu 2011). Strullu-Derrien et al. (2014a) examined the hydraulic properties of a very early wood with secondary thickening. Rudall and Bateman (2019b) discuss the cellularization of the megagametophyte, noting that an initial free-nuclear phase is found only in heterosporous lycophytes and in seed plants (including some fossils), also heterosporous, but it appears not to occur in heterosporous ferns.

Ecology & Physiology. Proctor (2014: p. 66) emphasized that endohydrous and homoiohydrous vascular plants clearly did not evolve as such (c.f. the birth of Athena), but that elements of the "vascular-plant package" were to be found in ecto- and poikilohydrous plants. Indeed, the distinction between these two "strategies" is not that sharp, and some ferns, for instance, may be homoiohydrous for only part of the year (e.g. Lösch et al. 2007), or gametophytes may be poikilohydric, sporophytes homoiohydrous.

Edwards (1993; Edwards & Richardson 2004; also Niklas 2015 for some cautionary comments) reviewed the anatomy of the vegetative parts of sporophytes of early land plants. For the evolution of xylem and the relation between details of cell anatomy and cell wall chemistry and water transport, plant support, etc., see Sperry (2003 and references); lignin may initially helped in keeping tracheary elements open, only later becoming involved in plant support. The formation of borate cross-linked rhamnogalactan II, at most sparse in bryophytes but present throughout extant vascular plants where it cross-links primary cell wall pectins, seems to have been an important early step in the development of lignified cell walls, and such rhamnogalactans are found in both fern sporophytes and gametophytes (Matsunaga et al. 2004). Cichan (1986) and Wilson (2015) and references relate details of xylem anatomy and water transport, linking extinct and extant plants; woods of early vascular plants may have been quite efficient in conducting water, although susceptible to embolism/cavitation (see Schenk et al. 2017 for the role of lipid surfactants in the xylem in preventing embolisms). Tracheophytes all have some kind of roots, although these have almost certainly evolved more than once (e.g. Raven & Edwards 2001; Pires & Dolan 2012), indeed, structures very like quiescent centres have evolved more than one (Fujinami et al. 2017). Tomescu et al. (2014) and Proctor (2014) suggest that roots would allow plant size to increase by providing stability and facilitating increased water uptake. For the evolution of root hairs, see the discussion above.

The function of stomata in extant vascular plants is simple, even if details of stomatal control are somewhat unclear: Stomata are involved in gas exchange. This may be a derived feature that can be placed somewhere in stem-group tracheophytes, hence the placement here. Stomata in fossils of early land plants are often large, examples being scattered regularly throughout the 50-140 μm length range (includes Zosterophyllum, Rhynia and Astroxylon), and in some Silurian plants the outer periclinal wall may be incomplete and in Devonian examples it is often thin, furthermore, stomata are not always found on parts of the plant where they are expected to be, i.e., in areas where there is likely to be much chlorophyllous tissue (Edwards et al. 1998; for stomatal pore length, see also Beerling & Woodward 1997). The discussion of the function of these stomata - some of the plants had vascular tissue of one sort or another - focusses on photosynthesis but also water uptake/generation of a transpiration stream so improving the supply of nutrients to the sporangium (e.g. Edwards et al. 1998; Haig 2013). Stomata are also known from gametophytic plants (Edwards et al. 1998), and assuming their function there was the same as that in the sporophyte, sporangial drying cannot be involved. Renzaglia et al. (2017) note that the Lower Devonian fossils Sporogonites and Tortilicaulis (relationships uncertain, perhaps polysporangiophytes) seem to have collapsing stomata similar to those in hornworts, and stomata on capsules of bryophytes are involved in the drying out of the capsule and spore discharge. If sporangial drying is indeed the original function of stomata (see also McAdam & Brodribb 2012a, b), then the central role that stomata now play in photosynthesis in vascular plants becomes a spectacular example of an exaption.

Initially it seemed that the response of photosynthesis to red light and passive stomatal control of leaf hydration were best tagged to this node (see above: McAdam & Brodribb 2011). It was thought that the mechanism of stomatal closure in ferns was like that of lycophytes rather than that of seed plants (McAdam & Brodribb 2011, 2012, 2013; see also Haworth et al. 2011, 2013; McAdam et al. 2016); it is passive, and abscisic acid is not immediately involved. However, some recent work suggests that in some ferns, at least, there was an active response to both abscisic acid and CO2, the nature of the response depending both on the species involved and growth conditions, e.g., humidity (Hõrak et al. 2017; c.f. Brodribb & McAdam 2017). Note that there seems to be no movement of the stomatal aperture in Equisetum, where it is permanently closed, and this may also be the case in Psilotum - although there is movement in Lycopodiales (Cullen & Rudall 2016; Roelfsema & Hedrich 2016). Indeed, in ferns, at least, part of the pathway involved in stomatal closing in seed plants is involved in spore dormancy (as perhaps in bryophytes) and in the determination of the sex of the gametophyte, and there, too, there is antagonism between gibberellic acid and abscisic acid (McAdam et al. 2016; see Cullen & Rudall 2016; Chater et al. 2017 and references for the evolution of stomatal patterning and development). However, stomata in vascular plants are clearly involved in gas exchange (for further discussion see above)!

Kenrick et al. (2012) discussed the effect of vascular plants on the carbon cycle. There are broad correlations between atmospheric CO2 concentration and stomatal size that have important implications for plant productivity, transpiration, and rock weathering. When the CO2 concentration of the atmosphere is low, leaves tend to have higher densities of smaller stomata, allowing more CO2 to diffuse into the leaf, when concentrations increase, relationships are the reverse (e.g. Franks & Beerling 2009). It has also been suggested that CO2 concentration, guard cell size, and genome size can be linked, the first and last being broadly correlated over the last 400 Ma (Haworth et al. 2011), although genome size reconstructions (see their Fig. 4) and some other aspects of this story (see also Franks et al. 2012) are difficult to understand. Within vascular plants there is a great range of relative pore area (pore area/total guard cell + pore area), that of Huperzia and Nephrolepis being much lower than that of flowering plants, and there is even greater variation in the rate of stomatal opening; in the two plants mentioned there was no movement of K+ ions that greatly affects the whole process (Franks & Farquhar 2006). For water use efficiency throughout the Phanerozoic, see Franks and Beerling (2009a) and Assouline and Or (2013: a different interpretation for higher CO2 concentrations). The carbon cycle was further affected by the evolution of vascular plants with well developed secondary thickening, whether producing mostly bark or both bark and wood; much carbon could be sequestered in these tissues (see elsewhere).

The epiphytic habit may have become established quite early on in some clades of vascular plants. Thus Hymenophyllum and part of Trichomanes, commonly epiphytic on tree ferns, a relatively old clade (Schuettpelz 2007; Hennequin et al. 2008; see also Schuettpelz & Pryer 2009), diversified some time between the early Middle Jurassic and the Late Cretaceous, indeed, epiphytes in Hymenophyllaceae are often low epiphytes, growing on the trunks and lower branches, perhaps an older habitat that that of crown epiphytes, which grow higher up in the tree and on branches (see also Lehnert et al. 2017; Lehnert & Krug 2019). About half - 190/380 species - of clubmosses, Lycopodiaceae, are epiphytic, and their diversification may have begun in the Late Cretaceous (Wikström & Kenrick 1997, 2001; Wikström 2001), however, Testo et al. (2018b) suggest that Phlegmariurus, a major clade, was ancestrally epiphytic (see also Field et al. 2016), and was perhaps growing in this habitat as early as the middle Jurasiic. Botryopteris (Ophioglossaceae) may have been another early epiphyte, apparently growing on the extinct marattialean tree fern Psaronius (Rothwell 1991).

One of the ways in which ferns and lycophytes, whether gametophyte or sporophyte, can grow in drier conditions is by being dessication tolerant, cutting down water loss, etc., even if the mechanism of stomatal control is overall similar to that in ferns growing in more mesic conditions (McAdam & Brodribb 2013).

Ferns, gymnosperms and lycophytes tolerate nutrient-poor (but sometimes rich in non-essential minerals) conditions, perhaps the ancestral conditions for many of them (Page 2004). The evolution of some genes in plants involved in mycorrhizal associations that facilitate phosphorus uptake by the plant may be pegged to this node (Delaux et al. 2015). Note, however, that any advantages to the plants in which mycorrrhizae were first established may differ from those to extant plants (Maherali et al. 2016).

Plant-Animal Interactions. Kato (2017) summarises a number of plant-animal interactions. Lutzoni et al. (2018) discuss in detail the possible association of diversification in the species-rich ascomycete group Leotiomyceta and the early diversification of vascular plants, including the origin of spermatophytes.

Bacterial/Fungal Associations. Early vascular plants are likely to have had a variety of associations with fungi (e.g. see Rimington et al. 2014). Mycorrhizal associations with AM Glomus are common in extant vascular plants. The commonest mycorrhizal association seems to be the Paris type where the hyphae are intracellular, forming coiled structures between the plant cells (F. A. Smith & Smith 1997; Winther & Friedman 2008), although the Arum type, with extensive intercellular hyphal growth and intracellular arbuscules and vesicles is also frequent (Field et al. 2015d) and that is the association best understood at the molecular level (Cosme et al. 2018) - and although we talk about mycorrhizal "types", there is more a continuum of variation (Dickson 2004; Dickson et al. 2007). In Ophioglossum and Botrychium, some Lycopodiaceae, Psilotum, and a few ferns and angiosperms, etc., associations are variously with the echlorophyllous gametophytic and/or sporophytic stages (Winther & Friedman 2007, 2008, 2009; Hynson et al. 2013). Around 10% of all vascular plants are mycoheterotrophic for all or part of their life cycles (Leake & Cameron 2010), a number driven by the >26,000 species of Orchidaceae nearly all of which have such associations when they are germinating; there are perhaps 1,000 mycoheterotrophic species in the Lycopodiales and Monilophytes combined (Winther & Friedman 200). Full mycoheterotrophy, known from ca 880 species (this number includes ferns, etc., with echlorophyllous gametophytes), has evolved 46 or more times, including in liverworts (Merckx et al. 2013a). For "ancestral" AM genes in Selaginella, see Bravo et al. (2016).

Genes & Genomes. There cannot be a genome duplication common to all tracheophytes since there is no evidence of genome duplication in Selaginellaceae (e.g. Banks et al. 2011; Baniaga et al. 2016). A PCA analysis of functional protein domains suggested that Selaginella was not that different from the few gymnosperms in the study, but together they were rather different from most angiosperms (Wan et al. 2018).

Chromosome numbers in heterosporous vascular plants are lower and genome sizes smaller than in homosporous plants. Thus in pteridophytes (including lycophytes) n = 57 on average, but in heterosporous water ferns n = 13.7, in angiosperms n = 16, in homosporous Lycopodium chromosome numbers are higher than in heterosporous Isoëtes and Selaginella, and chromosome numbers in gymnosperms are on average also low (Barker 2013; see also Sessa & Der 2016) - not that although some flowering plants in particular have high chromosome numbers, the average is low. In ferns, genome duplication followed by gene silencing but not chromosome loss may be responsible for the high numbers (Haufler 1987; Barker 2013; Sessa & Der 2016 and references), while in angiosperms in particular there is extensive gene and genome remodelling after polyploidization events (see elsewhere). It is suggested that there was a genome duplication in the ancestor of the [Salviniales [Cyatheales + Polypodiales]] clade (F.-W. Li et al. 2018).

Three new families of transcription-associated proteins may have evolved in this general area (Lang et al. 2010: hornworts not included; see also Zhu et al. 2012, Lang et al. not cited). For genome sizes in monilophytes and lycophytes, see Nakazato et al. (2008) and Leitch and Leitch (2013), while Baniaga et al. (2016) also discuss the rate of genome size evolution in various clades of vascular plants - generally low, but very low in Selaginella, quite high in ferns and still higher in many angiosperms.

For the evolution of the chloroplast genome in other than seed plants, see Wolf and Karol (2012). Chloroplast genomes seem particularly labile in many taxa ouside the angiosperms (Guisinger et al. 2011 for references), although less so in Lycopodiaceae where they seem to be highly conserved (Mower et al. 2018).

The order of genes in mitchondria is relatively invariable in bryophytes but much more variable in vascular plants (Y. Liu et al. 2014a).

Chemistry, Morphology, etc.. Condensed tannins are polymerized in a chloroplast thylakoid-derived tannosome (Brillouet et al. 2013). In extant vascular plants, lignins are rich in guaiacyl units (Harris 2005), and the evolution of the cinnamyl/sinapyl alcoholase gene family involved in the synthesis of the hydroxycinnamyl alcohol monomer units (p-coumaryl/p-hydroxyphenyl (H), guaiacyl (G), and syringyl (S) units) that ultimately constitute lignin can perhaps be pegged to this node (Guo et al. 2010; c.f. Gómez-Ros et al. 2007; Zu et al. 2009), although the first two are found in Physcomitrella, at least (see also above). Some xylans may be restricted to the sporophyte (Eeckhout et al. 2014).

Programmed cell death in vascular plants is involved in tracheid development, for instance (see van Hautegem et al. 2015). Edwards (2003) and Edwards et al. (2003) examined conducting cells of early tracheophytes and compared the morphologies of the cells involved with those of the bryophytes s.l.. For a general discussion on the evolution of water-conducting cells, with particular attention to wall sculpturing and its nature, see especially Kenrick and Crane (1991, 1997), Cook and Friedman (1997), Friedman and Cook (2000) and Edwards et al. (2006). However, understanding the fossil record is difficult in part because our knowledge of the development and nature of the wall thickening even of extant vascular plants is surprisingly poor, and exactly where in the wall lignin is deposited affects water conductance (Sperry 2003). The developmental control of conducting tissue in moss gametophytes and vascular plant sporophytes suggests that at one level they may not be that much different (see also above). For a general comparison of tracheary cells, see Bailey and Tupper (1918).

It is unclear exactly when/in what clade an endodermis evolves (Raven & Edwards 2001: p. 385).

For details of the sporophyte-gametophyte boundary, see Duckett and Ligrone (2003).

Kaplan (1997; vol. 3, 2001) provides an extensive discussion and analysis of the basic morphology of lycophytes and monilophytes in particular.

Phylogeny. The relationships [lycophytes [monilophytes + lignophytes]] are very commonly found, although not in some analyses of chloroplast genomes in Ruhfel et al. (2014) or in the mostly chloroplast gene analysis of Z.-D. Chen et al. (2016), while in de la Torre-Bárcena et al. (2009: expressed sequence tags) a clade [monilophytes [lycophytes [bryophytes s.l.]]] was shown as sister to seed plants, although relationships among seed plants was the major focus there.

Classification. Lycophytes and monilophytes or ferns have traditionally been placed together in the pteridophytes.

IIA. LYCOPODIOPSIDA Bartling / LYCOPHYTES

Root-bearing stems from angles of branches, roots lacking cuticle [?true], branching exogenous, dichotomous, protoxylem endarch, fine roots plagiotropic; mitosis monoplastidic; stem with protostele [= actinostele], protoxylem exarch, endodermis +; leaves small, with a single vein, phloem surrounding xylem; sporangia lateral, in strobili, 1/leaf, adaxial, often heart-shaped, dorsiventrally flattened, dehiscence transverse, along line of conspicuously thickened cells; MTOC on nuclear membrane; zygote with variable plane of first cell division, elongating, with quadrant/octant formation, shoot developing towards the archegonial neck [from hypobasal cell, endoscopic], embryonic axis reorients during development, root lateral with respect to its longitudinal axis [plant homorhizic]; nuclear genome [1C] (78-)1,165(-11,704) Mb; mitochondrion with loss/pseudogenisation of 6 genes [inc. all 4 ccm genes], some group II introns lost. - 3 families, 17 genera, 1,340 species.

Age. Larsén and Rydin (2015) estimated an age of about 407 Ma, Magallón et al. (2013) an age of around 383.7 Ma, Laenen et al. (2014) an age of ca 303.1 Ma, B. Zhong et al. (2014b) an age of (403-)386.3(-377.4) Ma, Villarreal and Renner (2014) an age of only around 270 Ma but Evkaikina et al. (2017) an age of ca 376 Ma for crown-group Lycopodiopsida and Testo et al. (2018b) an age of (414-)403.3(-394) Ma; see also P. Soltis et al. (2002).

Evolution: Divergence & Distribution. For the early evolution of Lycopodiopsida, which have a rich fossil record, see Gensel and Berry (2001), Wellman et al. (2009), Ambrose (2013), Gerienne et al. (2016 and references) and Xue et al. (2016: Drepanophycus rhizomes ca 410 My). they include zosterophylls, which have circinately-coiled stems and sporangia in two rows (see also Gensel 1992; Kenrick & Crane 1997; Gonez & Gerienne 2010). Tree lycopsids are known from early late Devonian deposits ca 380 Ma (Berry & Marshall 2015).

Much work has been carried out on fossil lycopodiopsids, known from the Devonian onwards, a number of which were tall tree of Carboniferous forests. Heterospory developed early (see Phillips 1979) for a good example), and such trees are in the same immediate clade as today's lowly Isoëtes. Zosterophylls and some other genera may form a clade with Lycopodiaceae (Gomez & Gerienne 2010).

Laenen et al. (2014) give some crown-group ages for genera.

Ecology & Physiology. Green (2010) described photosynthesis in this clade, focussing on the tree-like extinct lycopsids and the extant Isoëtaceae. Tree lycopsids dominated in late Carboniferous peat swamps, but are known well before, e.g. from early late Devonian deposits ca 380 Ma in forests of Protolepidendropsis puchra in Svalbard, were were made up of trunks up to 4 m tall and 9 cm across, although the flared base was up to 20 cm across (Berry & Marshall 2015: evidence for secondary thickening indirect). When tree lycopsids were flourishing, from the end of the Carboniferous to the beginning of the Permian, a period of about 100 Ma, it was a time when atmospheric CO2 concentrations were low and oxygen concentrations were high. Air canals/aerenchyma permeated both above- and below-ground parts of the plants, i.e. they had a parichnos system. This enabled oxygen to move to the roots, where it was used up in respiration, while CO2 produced by the plant remained in the plant and also moved into the plant from the anoxic CO2-rich soil, ultimately going to the leaves, where it was used up in photosynthesis; this is the lycopsid photosynthetic pathway, LPP (Green 2010, esp. p. 2258). However, although movement of oxygen within the plant may have been facilitated, a pad of tissue at the base of the rootlets cut off their air spaces from those of the stigmarian roots, yet overall CO2 movement and photosynthesis in the plant may have been little affected (Boyce & DiMichele 2015).

Some lycopsids reached perhaps 50 m tall and 2 m d.b.h., but they had little secondary or even primary phloem. Cork and a little xylem were all that was produced by any cambium, and the plants seem to have undergone a period of establishment growth (c.f. palms) when the stigmarian root system became established and the apical meristem of the stem enlarged so that when the stem finally began to elongate, it produced tissue of the "adult" thickness (e.g. Phillips & DiMichele 1992). The phloem, in which the sieve cells may have had degenerate nuclei, is likely to have been very long lived (Boyce & DiMichele 2015).

The rootlets on the Carboniferous stigmarian roots of tree lycopsids have the same branching pattern as in extant Isoëtes; branching is dichotomous, the roots become narrower at each dichotomy, but there is otherwise no secondary thickening/tapering, and there are root hairs - add similarities in anatomy, and Isoëtes is practically identical (Hetherington et al. 2016a). From the analysis of vascular tissue, it seems that auxin travelled from the base to the apex of the stigmarian root, the "wrong" direction for a stem - which it is morphologically - but the "right" direction for a root - which it is functionally (Sanders et al. 2011; see also Menand et al. 2007 for a similar example); the bipolar roots of seed plants show a similar pattern of auxin flow, but roots there are independently evolved (Kenrick 2013). Hetherington and Dolan (2017) observed that roots of all three extant groups of lycopsids - and their fossil relatives, where known - were morphologically very similar, having root hairs, root caps (Friedman et al. 2004, q.v. for caveats over roots in tracheophytes in general, as in Hetherington et al. 2016b; Hetherington & Dolan 2018), but were borne on a great variety of structures on extinct and even extant lycopsids; they did not rule out the possibility that these roots had evolved independently in Isoëtaceae (see below) and the other lycopsids.

Bacterial/Fungal Associations. There are various associations of fungi with the sporophytes of Lycopodiopsida, although such associations may quite often be absent (Rimington et al. 2014, 2017; Kenrick & Strullu-Derrien 2014; c.f. Winther & Friedman 2008). This absence may be connected with the thinness of the roots, which are often less than 1 mm across in lycopods and Selaginella, and mycorrhizae are uncommon in such roots (Pressel et al. 2016).

Vegetative Variation. It has been suggested that root-bearing axes of many taxa in this group are modified dichotomizing branches (stigmarian roots), the rootlets/ultimate roots being modified leaves (Pigg 1992; Rothwell & Erwin 1985; Rothwell 1995; Kenrick 2013) or organs sui generis (Matsunaga & Tomescu 2016; see also Tomescu et al. 2014). Crane and Kenrick (1997) thought that all appendicular structures in Lycopodiopsida - microphylls, ligules, sporangia, and stigmarian roots - were equivalent, i.e., a position consistent with the former hypothesis. In Isoëtes the roots are exogenous (Rothwell & Erwin 1985) and this is also true of the K-branching that produces root-bearing axes in other Lycopodiopsida, as in a drepanophycalean lycophyte (i.e. in the same immediate clade as Lycopodium); in such K-branching, one branch of a dichotomy becomes a root-bearing axis and may have a few leaves along its upper part, the other becomes an erect leafy stem (Matsunaga & Tomescu 2016). Rothwell and Erwin (1965) described the roots of Lycopodium and Selaginella as being adventitious, while those of Isoëtes abscise in much the same way as microphylls (Hetherington & Dolan 2017 for references). The small, dichotomising ultimate roots (Matsunaga & Tomescu 2016) may have a root cap, and this is reported from other Lycopodiopsida and lycopsids (Hetherington & Dolan 2017). However, Hetherington and Dolan (2018) have recently described the apices of rooting axes of Asteroxylon, a ca 407 Ma lycopsid from the Rhynie chert - no root cap was evident, just a continuous epidermis. This suggested that the root cap - and associated roots - had evolved at least twice, one here and once in the seed plant (?or euphyllophyte) clade (see also Hetherington et al. 2016b).

Genes & Genomes. Leitch and Leitch (2013) found that Lycopodiopsida had rather small nuclear 1C genome sizes - (0.086-)1.7(11.96) pg - but sampling was poor; see Sessa & Der 2016 for size in the context of homo-/heterospory).

The chloroplasts of Lycopodiaceae and Isoëtaceae have 118-122 genes and 21-22 introns (Mower et al. 2018).

For the loss of various mitochondrial genes and the distribution of introns - ca 11 cis-splicing introns seem to have been gained - in this clade, see W. Guo et al. (2016b). The ccm genes that have been lost represent the loss of the mitochondrial cytochrome c maturation pathway.

Chemistry, Morphology, etc.. Selaginellales in particular, but also Lycopodiales, may contain fair amounts of silica (Trembath-Reichert et el. 2015).

Stubblefield and Rothwell (1981) discuss embryogenesis in Lycopodiopsida.

Meiosis is poyplastidic and anastral [??].

For general information, see also Kenrick and Crane (1997) Boyce (2005a) and Ranker and Haufler (2008), for fossil members, see Gensel et al. (2013), for sporophytes and their placentae, see Hilger et al. (2002), for details of phloem, see Evert (1990a), and for meiosis, MTOCs, etc., see Brown and Lemmon (2008).

Classification. Lycophytina include the three families below and their fossil relatives, and these latter include zosterophylls (e.g. Gensel 1992; Kenrick & Crane 1997; Gonez & Gerienne 2010; Gerienne et al. 2016 and references). See The Pteridophyte Phylogeny Group (2016) for a classification, also Christenhusz et al. (2011a) and H.-M. Liu (2016). Each family is sometimes put in a monofamilial order...

LYCOPODIACEAE P. Beauvois  - Back to Lycopodiopsida

Plant terrestrial or epiphytic; (roots with glomeromycotes and/or mucoromycotes); root apical meristem 3-4-tiered, open, with common initial zone/quiescent centre, or closed, protoderm, root cap and ground meristem initials all separate, endodermis 0; shoot apical meristem complex, plasmodesmatal density in whole SAM 0.4-4.2[mean]/μm2; (strobili 0); pollen tapetum 0; (meiosis polyplastidic); (gametophyte mycoheterotrophic), basidiomycetes, glomeromycotes +; male gametes with 2-20 cilia; n = 34 [quite often; lots of other numbers], nuclear genome [1C] ca 3.76 pg.

15 [list: to subfamilies]/390: Phlegmariurus (250), Lycopodiella (60). World-wide, esp. South America>

Age. The age of crown-group Lycopodiaceae is about 265 Ma (Laenen et al. 2014: crown age of Huperzia), 167 Ma (Larsén & Rydin 2015) or (383-)368.4(-325 Ma (Testo et al. 2018b).

Hup: (223-)199.5(-175) Ma (Testo et al. 2018b), Ly-Lell, (320-)293.6(-262) Testo et al. 2018b), Ly (236-)210.4(-202) Ma (Testo et al. 2018b) and Lell ca 118.8 (Testo et al. 2018b).

Evolution: Divergence & Distribution. The initial distribution patterns developed in Lycopodiaceae reflect vicariance events associated with the break-up of Pangaea (Testo et al. 2018b).

Wikström and Kenrick (1997, 2001) and Wikström (2001) discuss diversification and phylogeny of extant Lycopodium s.l. and relatives. Although diversification in Lycopodium s.l. may have begun some 200 or more Ma, the age of fossils with the distinctive plectostele that characterises Lycopodium s. str., most páramo species are probably derived from epiphytic species within the last 15 Ma (Wikström et al. 1999; Sklenár et al. 2011). Indeed, ca 60/150 of the Neotropical species of Phlegmariurus are terrestrial, growing in alpine grassland (the clade is derived from epiphytes in montane forests) and has diversified there in the last 10 Ma or so (Testo et al. 2019).

Ecology & Physiology. About two thirds of the species of Huperzia, a number of species of Phlegmariurus, and all told around 200 species in the family, are epiphytic (Wikström et al. 1999; Schuettpelz & Pryer 2009; Zotz 2013). Phlegmariurus in particular shows extreme ecological flexibility (Testo et al. 2018a).

Renzaglia and Whittier (2013) discuss the intereractions of the gametophyte and young sporophyte when the former is subterranean and mycoheterotrophic.

A member of Sebacinales-Sebacinaceae from Diphasiastrum alpinum was also found on Calluna vulgaris growing nearby, and nutrients moved from Calluna to the echlorophyllous lycopod gametophyte (Horn et al. 2013; c.f. Rimington et al. 2016).

9/10 species of Lycopodiaceae studied were aluminium accumulators with >1000 mg Al kg-1 (Schmitt et al. 2017).

Bacterial/Fungal Associations. Glomales are associated with the echlorophyllous gametophytic and subterranean sporophytic stages in some lycophytes (Merckx et al. 2013a; Imhof et al. 2013), and these fungi may also be associated with the autotrophic sporophyts (Winther & Friedman 2008). The Glomus involved in the mycoheterotrophic association forms a clade with "species" involved in similar associations in Arachnitis (Corsiaceae) and Botrychium (Ophioglossaceae) (Winther & Friedman 2007, 2008); together, they are part of the Glomus A group (Schußler et al. 2001). For other fungal associates of Lycopodiaceae, see Rimington et al. (2014), Lehnert et al. (2016) and Pressel et al. (2016); the latter note that different clades of mucoromycote may be found in the same plant in some Lycopodiaceae.

Genes & Genomes. The trnL and trnS genes in the mitochondrial genome of Huperzia, but not in those of other Lycopodiaceae, may have come from chlamydial bacteria by lateral gene transfer (Knie et al. 2015).

Chemistry, Morphology, etc.. For general information, see Øllgaard (1990), for spermatogenesis, see Renzaglia et al. (1994b).

Phylogeny. For relationships in the family, see Field et al. (2016) and Burnard et al. (2016); Phlegmariurus, along with Phylloglossum and Huperzia, form a clade sister to the rest of the family. The bulbiferous Huperzia selago group is sister to the rest of the genus, which is mostly epiphytic; there are New and Old World clades (Wikström et al. 1999). Testo et al. (2018a) discussed relationships within the morphologically and ecologically variable New World species of Phlegmariurus.

Classification. For the current classification of the group, see The Pteridophyte Phylogeny Group (2016); it is very different from the classification in account by Øllgaard (1987), but c.f. Øllgaard (2016).

[Isoëtaceae + Selaginellaceae]: leaves with adaxial ligule; plant heterosporous; megaspore wall with much silica, outer and inner exine separated by discontinuity; gametophyte development endosporic, with initial free-nuclear phase, intraplacental space +, transfer cells 0; nuclear genome [1C] ca 0.27 pg; mitochondrion with loss of 15 genes.

Age. The age of this node is around 209 Ma (Laenen et al. 2014), (386-)375(-360) Ma (Pereira et al. 2017), 381 Ma (Larsén & Rydin 2015) or (393-)383(-374) Ma (Klaus et al. 2017).

Evolution: Divergence & Distribution. Blackmore et al. (2000) score the cells of this pair as not being monoplastidic, but c.f. Brown and Lemmon (1990), Schuette and Renzaglia (2010), etc.. For spore wall ultrastructure in fossil members of this clade (homosporous!), see e.g. Wellman et al. (2009).

Genes & Genomes. W. Guo et al. (2016b) found that a total of 21/46 mitochondrial genes were absent from or pseudogenised in Isoëtes and Selaginella.

Chemistry, Morphology, etc.. For cellularization in early gametophyte development, see Rudall and Bateman (2019b), and for embryo development, see Renzaglia and Whittier (2013).

ISOËTACEAE Dumortier  - Back to Lycopodiopsida

Plant herbaceous, terrestrial or aquatic, evergreen or variously deciduous; mycorrhizae at most uncommon; roots arising from beneath the corm/rhizome, 2-3-tiered, meristem closed, protoderm and root cap from common initials, with a central air space, vascular bundle single, excentric, xylem abaxial to phloem; stem cormose, unbranched, 3(-2)-lobed, (elongated, dichotomously branching - I. andicola); SAM with mean plasmodesmatal density 2.2-4.1/μm2 [cell interface-specific plasmodesmatal network]; xylem mesarch; vascular cambium +; stomata 0 (+); leaves with several vascular strands; megasporangia indehiscent, decaying, trabeculate; megaspores 50-300/megasporangium, surface tuberculate; microsporogenesis successive, tetrads decussate, blepharoplast branched; (microspores monolete); tapetum glandular; sporophyte-gametophyte junction with dead gametophytic cells, male gametes with many cilia [10-20]; embryo lacking suspensor; n = (10) 11.

1 [list]/ca 250. More or less world-wide.

Age. The age of crown-group Isoëtes is (154-)147(-145) Ma (Pereira et al. 2017), (235-)165, 147(-96) Ma (Larsén & Rydin 2015) or ca 251 Ma (C. Kim & Choi 2016).

Evolution: Divergence & Distribution. For ages in Isoëtes, see also Pereira et al. (2017).

For the evolution of the whole group, including arborescent members and fossils that have been linked with Isoëtes, see Retallack (1997c), Grauvogel-Stamm and Lugardon (2001), Pigg (1992, 2001) and above; Pleuromeia may not have been the "ancestor" of Isoëtes.

Ecology & Physiology. Some species of Isoëtes take up CO2 from the mud in which they grow via their very well developed roots, Stylites (= I. andicola) and submerged individuals of other species of Isoëtes even lacking stomata - I. andicola never has stomata (Bristow 1975; Keeley 1998; Raven et al. 1998 and references). Photosynthesis is by a sort of modified CAM pathway, the lycopsid photosynthetic pathway (see above), and this has a substantially more ancient origin than does the CAM pathway in flowering plants (Edwards & Ogburn 2012).

Corms of some species of Isoëtes can tolerate extreme dessication (Proctor & Tuba 2002; Gaff & Oliver 2013). For Late Embryogenesis Abundant (LEA) genes and gene families, involved in dessication tolerance in land plants in general, see Artur et al. (2018).

Bacterial/Fungal Associations. Mycorrhizae would not be expected in a largely aquatic group, but arbuscular mycorrhizae (and dark septate hyphae) were found in some roots of two species of Isoëtes in a lake in the Czech Republic, although their identity and what they might have been doing in the plant is unclear (Sudová et al. 2011).

Genes & Genomes. The chloroplast IR seems to have been converted into a direct repeat, but then became inverted again (Mower et al. 2018).

Chemistry, Morphology, etc.. The growth and anatomy of Isoëtes is poorly understood (e.g. Gifford & Foster 1988). Doyle (2013) described the vascular cambium as producing both xylem and phloem to the inside and parenchyma to the outside (c.f. Kaplan 1997, vol. 3: chap. 19).

Strobili are not obvious; all leaves may be fertile. Isoëtes, at least, has placental cells with thickened, nacreous walls.

For general information, see also Jermy (1990); for megaspore morphology, quite diverse, see Hickey (1986).

Phylogeny. For relationships within Isoëtes, see Rydin and Wikström (2002), Hoot et al. (2006: ?rooting), Choi et al. (2018: East Asian taxa) and especially Larsén and Rydin (2015). A clade widely distributed on Gondwanan continents is sister to the rest (or [African taxa [Australian, S.E. Asian, Central Americacn taxa + the rest]] - Choi et al. 2018), and overall there is a fair bit of geographical signal in the relationships (Larsén & Rydin 2015, q.v. for dates, etc.; Pereira et al. 2017).

SELAGINELLACEAE Willkommen  - Back to Lycopodiopsida

Plant usu. terrestrial; (roots with mucoromycotina); SiO2 accumulation common; syringyl lignin +; rhizophores +, growing in length via intercalary meristems/0, root with apical cell, closed [protoderm, root cap both separate], with a single apical cell, hypodermis suberized/with Casparian strip; stele in central cavity surrounded by trabeculate endodermal cells; (vessels +); stem with single apical cell, plasmodesmatal density in whole SAM 27-44[mean]/μm2 [cell lineage-specific plasmodesmatal network]; stem mono-(-4-)stelic, ± actinostelic); leaves (often 4-ranked), (often anisophyllou; sporophylls (often 4-ranked); sporangia ± spherical, megaspores 4/megasporangium, surface with solitary protrusions (scabrate, reticulate); microspores (in tetrads), often echinate; n = (7-)9(10, 12), nuclear genome [1C] 0.08-0.19 pg; b class="apo">plastomes with 68-93 genes

and b class="apo">7-11 introns

, b class="apo">very high GC content

; trans splicing of some mitochondrial group II introns.

1 [list]/ca 700. World-wide.

Age. The crown-group age of this clade is (352-)322, 312(-310) Ma (Arrigo et al. 2013), ca 328 Ma (Larsén & Rydin 2015) or (387-)373(-354) Ma (Klaus et al. 2017).

Fossils (Selaginellites resimus) from the Lower Carboniferous-Visean ca 350-333 Ma are the earliest Selaginellaceae (Rowe 1988).

Evolution: Divergence & Distribution. For ages of clades within Selaginella and an evaluation of the fossils record there, see Klaus et al. (2017).

Klaus et al. (2017) provide details of the evolution and distribution of Selaginella; species here are fairly old compared with those in Cycadales and Pinales. Petersen and Burd (2018) discuss heterospory in the genus, perhaps a particular advantage in more shaded habitats; species growing in such conditions tend to have larger meagspores than those growing in open habitats.

The small subgenus Selaginella, sister to the rest of the genus, is relatively undistinguished vegetatively, having monomorphic sterile leaves but lacking rhizophores, and subgenus Boreoselaginella also has more or less monomorphic sterile leaves. Details of their morphology, especially those of subgenus Selaginella, will affect character polarization. Species in subgenus Selaginella lack a rhizophore, but rhizophore-like organs are found in fossils (e.g. Matsunaga et al. 2017); note that the earliest fossils cannot be included in subgenus Selaginella (Klaus et al. 2017).

Ecology & Physiology. Selaginella grows in habitats varying from very humid and with considerable shade to open and xeric. Plants in the latter habitats reconstitute quite nicely when water is added (Pampurova & van Dijck 2014). Dessication tolerance has evolved several times (Korall & Kenrick 2004), a niche shift to drier conditions perhaps characterizing subgenus Tetragonostachys (= subgenus Rupestrae) as a whole (Arrigo et al. 2013). VanBuren et al. (2018) found great haplotype variation in the dessication-tolerant S. lepidophylla, and it has a small genome - around 109 Mb - that contrasts with the larger genomes of dessication-tolerant angiosperms (Farrant et al. 2015). However, Late Embryogenesis Abundant (LEA) genes are prominent, as in dessication tolerant angiosperms (see also Z. Xu et al. 2018; Artur et al. 2018), and there are also interesting tandem duplications of genes, also apparently involved in dessication tolerance (VanBuren et al. 2018a). Interestingly, S. tamariscina, also dessication tolerant, has a genome 301 Mb in size, at the other extreme for Selaginella (Z. Xu et al. 2018). Stem ages of clades now growing in more or less xeric habitats are late Permian/early Triassic, around 260-252 Ma, one of the clades involved, subgenus Pulviniella, being described as an example of extreme niche conservatism lasting hundreds of millions of year (Klaus et al. 2017).

In a comparison with ferns (6 species) growing in similar habitats on the forest floor in Costa Rica, the two groups had similar net photosynthesis and water use efficiency, but Selaginella (7 species) had a higher leaf mass/area and nitrogen use efficiency in photosynthesis, and shorter, more dense stomata (Campany et al. 2018).

Bacterial/Fungal Associations. For fungal associates of Selaginellaceae, see Rimington et al. (2014) and Lehnert et al. (2016).

Vegetative Variation. Angle meristems (dorsal and ventral) in Selaginella form in the axil of the bifurcation of the shoot apex and develop into rhizophores (Jernstedt et al. 1992). The rhizophore initially has a group of apical cells and is positively geotropic, but at the level of transcriptome analyses it differs from both shoot and root; genes associated with both, perhaps especially the root, are expressed, but also a set of unique genes (Mello et al. 2018). When the rhizophore divides it forms roots, the meristem soon comes to consist of a single cell, and root hairs and a root cap form when the root becomes subterranean (Lu & Jernstedt 1996); auxin transport is acropetal (c.f. seed plants: Matsunaga et al. 2017). Root elongation is distinctive, not proceeding by divisions of the immediate derivates of the apical cell but by the activity of an intercalary meristem that is some way behind the apex (Otreba & Gola 2011). See Mello et al. (2018) for the connection between auxin response and roots or root-like structures across land plants in general, and Damus et al. (1997) for the root hypodermis.

Genes & Genomes. There is no evidence of genome duplications in Selaginella (e.g. Banks et al. 2011; VanBuren et al. 2018a). Genome size has long been small in the clade, the range being around 100-144 Mb, and rather larger genomes above 1.3 pg seem to be derived; overall, the rate of genome size evolution is very low indeed, unlike the situation in groups like Brassicales, also with small genomes (Baniaga et al. 2016). Indeed, extremophiles - and several species of Selaginella qualify - in general (but not all - see Boea) tend to have small genomes (Baniaga et al. 2016).

Selaginella has a very highly reorganized chloroplast genome (Tsuji et al. 2007; Z. Xu et al. 2018; esp. Mower et al. 2018). Xu et al. (2018) found that the NDH genes of S. tamariscina, and some other genes, had been lost, although some of these other genes had been transferred to the nucleus, indeed, it has the fewest genes of any photosynthetic land plant, having 68-93 genes and 7-11 introns versus 118-122 genes and 21-22 introns in other lycophytes (Mower et al. 2018). Furthermore, the plastomes of Selaginella have the highest GC content of any land plant (D. R. Smith 2009), the IR is sometimes converted into a direct repeat (but it can become inverted again), and so on (Mower et al. 2018). The GC content of the genome is notably high (Y. Yu et al. 2019b).

The mitochondrial genome of Selaginella is small, but has the highest GC content of any land plant (Mower et al. 2018).

Chemistry, Morphology, etc.. For the synthesis of syringyl lignin in Selaginella, see J.-K. Weng et al. (2010); although also common in angiosperms, details of the synthetic pathway are quite different in the two (see also Harholt et al. 2016). Harholt et al. (2012) looked at the cell wall polysaccharides of Selaginella moellendorfii, comparing them with those of other land plants.

Korall and Taylor (2006: discussion about problematic quantitative characters, etc.), X.-M. Zhou et al. (2015b) and Zhou and Zhang (2015 and references) illustrate the sometimes baroque morphology of the micro- and megaspores of the genus. Hemsley et al. (1998 and references) describe how self-assembly is involved in putting together the exine.

Phylogeny. Korall and Kenrick (2004), Zhou et al. (2015a) and especially Weststrand and Korall (2016a) and Klaus et al. (2017) provide a comprehensive phylogeny of the genus. The Selaginella selaginoides clade (subgenus Selaginella) consistently comes out as sister to the rest of the genus, and the S. sanguinolenta clade (subgenus Boreoselaginella: Zhou & Zhang 2015) may be sister to the remainder (Zhou et al. 2015a; Klaus et al. 2017), although Weststrand and Korall (2016a) found that its position was uncertain and they included it in their large subgenus Stachygynandrum (Weststrand & Korall 2016b).

Classification. See X.-M. Zhou and Zhang (2015) and in particular Weststrand and Korall (2016b) for infrageneric classifications.

[MONILOPHYTA + LIGNOPHYTA] / EUPHYLLOPHYTA

Sporophyte growth ± monopodial, branching spiral; roots endomycorrhizal [with Glomeromycota], lateral roots +, endogenous; G-type tracheids +, with scalariform-bordered pits; leaves with apical/marginal growth, venation development basipetal, growth determinate; sporangium dehiscence by a single longitudinal slit; meiosis polyplastidic, MTOCs diffuse, perinuclear, migratory; blepharoplasts +, paired, with electron-dense material, centrioles on periphery, male gametes multiciliate; nuclear genome [1C] 7.6-10 pg [mode]; chloroplast long single copy ca 30kb inversion [from psbM to ycf2]; mitochondrion with loss of 4 genes, absence of numerous group II introns; LITTLE ZIPPER proteins +.

Note: In all node characterizations, boldface denotes a possible apomorphy, (....) denotes a feature the exact status of which in the clade is uncertain, [....] includes explanatory material; other text lists features found pretty much throughout the clade. Note that the precise node to which many characters, particularly the more cryptic ones, should be assigned is unclear. This is partly because homoplasy is very common, in addition, basic information for all too many characters is very incomplete, frequently coming from taxa well embedded in the clade of interest and so making the position of any putative apomorphy uncertain. Then there are the not-so-trivial issues of how character states are delimited and ancestral states are reconstructed (see above).

Age. The divergence of the monilophytes and lignophytes may date to 401-380 Ma (Leebens-Mack et al. 2005) or ca 380 Ma (Hennequin et al. 2008); Theißen et al. (2001) suggested an age of ca 400 Ma, Clarke et al. (2011: also other estimates) an age of (452-)434(-410) Ma and Magallón et al. (2013) estimated an age of around (422-)411-404(-394) My; (463.5-)428.9(-400.1) Ma are the ages in B. Zhong et al. (2014b), (478.4-)447.4(-415.2) Ma in Rothfels et al. (2015b), 455-427 Ma in Barba-Montoya et al. (2018) and 437.5-402 Ma in Morris et al. (2018). See also Pryer et al. (1995, 2000, 2001a, 2004), P. Soltis et al. (2002: a variety of ages, some very old), H. Schneider et al. (2002), Stein et al. (2012) and Larsén and Rydin (2015).

Evolution: Divergence & Distribution. For possible apomorphies of crown members of this clade, see Raubeson and Jansen (1992b), Kenrick and Crane (1997), and H. Schneider et al. (2009). Kenrick (2013) suggested that the development of an endodermis and the endogenous origin of root meristems may be associated with this node.

Extra-floral nectaries are scattered throughout [Monilophyta + Lignophyta], and Weber and Agrawal (2014) found that their evolution was often - but not always - associated with an increase in diversification of the clades in which they occurred; they indirectly facilitated diversification. Cyanogenic glycosides, α-hydroxynitrile glucosides that release hydrogen cyanide when acted upon by plant β-glucosides, are similarly scattered, and Bak et al. (2006) suggest they may be an apomorphy at this level. They note that ferns and gymnosperms have aromatic cyanogenic glucosides (derived from tyrosine or phenylalanine) and angiosperms have both aliphatic and aromatic glucosides (derived from leucine, isoleucine or valine), the latter group being derived - parallel evolution at different levels perhaps seems as likely.

The leaves of this clade, often called megaphylls or euphylls, vary considerably. The leaf supply in monilophytes seems to have evolved by dissection of an amphiphloic siphonostele, although the vascular system of the rhizome in some true ferns consists of sympodia (Karafit et al. 2005), while "leaf gaps" in seed plants are rather different and are associated with a stele that consists of a series of sympodia of collateral vascular strands (see Namboodiri & Beck 1968c; Slade 1971). Gunning et al. (1970) noted that xylem transfer cells seemed to be associated with leaf gaps, Equisetum, which had such cells, seeming to be an exception, but it does belong here; xylem transfer cells may be an apomorphy at this level, although transfer cells of one sort or another are quite widely distributed in land plants and some aquatic streptophytes. However, leaves with large blades in ferns and seed plants may be parallelisms rather than a synapomorphy, and so leaf gaps in the two may not be directly comparable; the term "megaphyll" can thus be used in a descriptive sense only (Namboodiri & Beck 1968c; Beck et al. 1982). Furthermore, the growth of fern leaves is accompanied by proliferation at the apex of the blade and that of angiosperms by proliferation at the base (Nardmann & Werr 2013), while Boyce (2005b, 2008a) noted the prevalence of marginal leaf growth in non-angiospermous leaf blades and of diffuse growth in angiosperm leaf blades - these may be oversimplifications, but they suggest differences between the two. Relating such differences to a common pattern of leaf development that includes that of the Lycopodiales/lycophytes is a challenge. For further details of the development/evolution of euphylls, see above.

Ecology & Physiology. Osborne et al. (2004) provide an ecological explanation for the origin of mega/euphylls based on falling CO2 levels in the latter part of the Devonian.

Support for the stem in monilophytes is provided largely by the lignified stereome, in the outer cortex, while the tracheids are relatively thin walled. In gymnosperms thick-walled tracheids provide much of this support, and in angiosperms, too, it is various elements of the vascular tissue that provide support (e.g. Rowe et al. 2004; Pitterman et al. 2011; Klepsch et al. 2015).

Genes & Genomes. For the absence of numerous group II introns at this node, see W. Guo et al. (2016b); depending of how characters are optimized, 12-21 introns may have been lost here. For high-level variation in transcription factors and transcription regulators, see F.-W. Li et al. (2018).

Literature on the substantial inversion of the chloroplast genome characterizing this clade is summarized by Szövényi (2016).

Chemistry, Morphology, etc.. For microtubules, see Schmit (2002). Kenrick and Strullu-Derrien (2014), Hetherington et al. (2016b), Hetherington and Dolan (2018) and others discuss roots and their evolution, which quite possibly happened independently in lycophytes. For lamina morphology and venation development, see Boyce (2005b). Noetinger et al. (2018) described the spore wall ultrastructure in Psilophyton dawsonii, noting that the wall consisted of inner and outer parts. The inner, probably lamellate but appearing homogeneous, was laid down by the spore, while the outer, consisting of a foveate basal part on which progressively smaller plates of sporopollenin were stacked, the tapering stacks ultimately forming spines, was of tapetal origin.

Phylogeny. Ferns and their relatives, the monilophytes or Polypodiopsida, and lignophytes, the extant members of which are seed plants or spermatophytes, are both well supported clades.

Classification. The clade [monilophytes + lignophytes] is sometimes called Euphyllophyta/the euphyllophytes.

IIB. POLYPODIOPSIDA Cronquist, Takhtajan & Zimmermann / MONILOPHYTA  - Back to Main Tree

Roots originating from the pericycle, lateral roots from the endodermis, apical meristem closed [protoderm and cap separate], with a single apical cell; stem with apical initial(s), plasmodesmatal density in whole SAM 19-56[mean]/μm2 [lineage-specific mitochondrial network]; stem with hypodermal and outer-cortical band of fibres [= stereome]; amphiphloic siphonostele +, discontinuities in stele in t.s. caused by frond gaps; protoxylem restricted to lobes of central xylem strand [giving a beaded appearance, hence monilophytes], xylem mesarch, tracheids with scalariform pits, G-type tracheids in protoxylem; phloem with refractive spherules in sieve tubes, phloem fibres rare; stem endodermis and pericycle +; leaves megaphyllous [ad/abaxial symmetry evolved first, then determinancy], development acropetal; petiole with multiple leaf traces coming from a U-shaped bundle; frond veins not anastomosing; sporangia grouped in sori, sporangium stalk 6< cells across, walls two cells thick, dehiscence by an exothecium, tapetum ± amoeboid, spores/sporangium 1000<, white, globose-tetrahedral, with orbicules, wall development centrifugal, exospore 3-layered, pseudoendospore +; gametophyte thalloid; antheridium wall ³5 cells thick, male gametes with 30-150 cilia, with numerous plastids and mitochondria; root lateral with respect to the longitudinal axis of the embryo [plant homorhizic]; x = ca 121, nuclear genome [1C] ca 14.3 pg/(0.25-)12(-148) Gb [or rather larger], chloroplast rps4 gene with nine-nucleotide insertion, LSC inversion 3.3 kb from trnG-GCC to trnT-GGU; loss of one group II mitochondrial intron.

Age. Oldest, at (482-)431(-420.6) Ma, is the estimate in Testo and Sundue (2016). Morris et al. (2018) estimated a crown-group age of 411.5-385 Ma for this clade, Magallón et al. (2013) estimated an age of around (404-)394.3-389.9(-382) Ma, ca 360 and ca 364 Ma are the ages in Schuettpelz and Pryer (2007) and Y-L. Qiu et al. (2007) respectively, (390.7-)368.5(-354) Ma in B. Zhong et al. (2014b), around 330 Ma in Villarreal and Renner (2014) and a Devonian age of around 370 Ma or considerably earlier is also consist with the findings of Elgorriaga et al. (2018); see also P. Soltis et al. (2002) for suggestions.

Evolution: Divergence & Distribution. Schuettpelz and Pryer (2009, esp. Tables 2, 3 in the Supplement), Rothfels et al. (2015b: Appendix S4) and Testo and Sundue (2016: e.g. Fig. S1) provide numerous ages for monilophyte clades, and for more ages, see also Y.-L. Qiu et al. (2007) and H. Schneider et al. (2004a). Schuettpelz and Pryer (2009) also list a number of fossil records (see also Rothwell & Stockey 2008).

Fossils associated with monilophytes very much expand one's idea of what a fern might be, even if one includes horsetails in that concept. Some of these early fossils are placed in Cladoxylopsida. Of these, the middle Devonian pseudosporochnalean Calamophyton, was a small, branched tree to 2.5(-4) m tall; the primary stem increased in width for up to 2 m height (to 20 cm diameter) and there was secondary growth towards the base of the stem/trunk. The primary stem, at up to 10 cm across, was very stout; the branches themselves branched dichotomously and bore small appendages/leaves and were shed as units (= cladoptosis) (Giesen & Berry 2013). Eospermatopteris was ca 1 m across at the base (Stein et al. 2007), while an individual of the cladoxylopsid Xinicaulis lignescens some 374 Ma (Devonian: Frasnian) ca 70 cm across showed diffuse secondary growth of the ground tissue and secondary thickening of the individual xylem strands (H.-H. Xu et al. 2017: secondary thickening compared to that of palms, it is "reasonably clear" (ibid., p. )that this is independent evolution).

Ferns as a whole shows little variation in estimates of disparity, that is, the amount/extent of morphological variation in a sample of taxa, over time (Oyston et al. 2016), although in leptosporangiate ferns in particular there has been a gradual increase. There have been several radiations of homosporous leptosporangiate ferns, the first in the Palaeozoic, giving rise to lineages that have since become extinct, and again in the Jurassic and the Cretaceous (Rothwell & Stockey 2008). General fern diversity decreased (along with that of the cycads) through the Cretaceous (Wing & Boucher 1998). Ferns appear to have temporarily dominated at least locally after the end-Cretaceous bolide impact/Deccan Traps eruptions (H. Schneider et al. 2004a). The eusporangiate Marattia and Angiopteris and the leptosporangiate royal and tree ferns (Osmundaceae, Cyatheaceae area) are almost living fossils and show little molecular and even morphological evolution (P. Soltis et al. 2002; B. Zhong et al 2014b; Rothfels et al. 2015b; Bomfleur et al. 2014a, c.f. H. Schneider et al. 2015).

Some 98% of living ferns are leptosporangiate ferns, and their main radiation may have have been after the diversification of the angiosperms in the late Cretaceous-early Caenozoic; initially it was a largely terrestrial radiation (Lovis 1977; H. Schneider et al. 2004a; Rothwell & Stockey 2008; Schuettpelz & Pryer 2009; H.-M. Liu et al. 2014). Well over 70% of ferns are eupolypods, and about one third of these are epiphytes, amd these make up ca 10% of all vascular epiphytes. Diversification of epiphytic ferns in particular occurred during the Palaeogene, perhaps linked with the Palaeocene-Eocene thermal maximum (H. Schneider et al. 2004a, b; Schuettpelz 2007; esp. Schuettpelz & Pryer 2009: Supplemental Tables 2, 3; Watkins et al. 2010; c.f. Dubuisson et al. 2009, in part). An exception is Trichomanes and relatives (but not Hymenophyllum and its relatives, for which, see Dubuisson et al. 2009) which diversified in the early Cretaceous; they are commonly epiphytic on tree ferns, which themselves had begun diversifying in the Jurassic (Schuettpelz 2007; also Schuettpelz & Pryer 2009; Rothwell & Stockey 2008 for early radiations of leptosporangiate ferns).

Testo and Sundue (2014) found that both terrestrial species and hemiepiphytes (where roots from the climbing plant establish a connection with the soil) were derived within a clade of epiphytic Polypodiaceae.

A provisional hierarchy of characters partly taken from A. R. Smith et al. (2006, 2008), Pryer et al. (1996) and Rothfels et al. (2012b) is given below. For a first stab at apomorphies for members of the polypod II clade, see Sundue and Rothfels (2013); their suggestions are largely followed here, and problems with sampling, character state delimitation, and different apomorphies produced by different methods of character optimization are taken into account as far as possible. Only unambiguous synapomorphies (Sundue & Rothfels 2013) are flagged (see also Pittermann et al. 2015 for petiolar vascular anatomy). Kaplan (1997, vol. 3: chap. 19) summarized the sporangium wall morphology of monilophyta; the annulus is an apomorphy of Polypodiidae, the leptosporangiate ferns (for annulus evolution, see H. Shen et al. 2017). For other possible apomorphies, see e.g. H. Schneider et al. (2009). More features may need to be added, e.g. if the megaphyllous leaves of most ferns and those of seed plants have evolved independently, which seems likely (see above); D. M. Wang et al. (2015) suggest that laminate leaves had evolved here by the Late Devonian (Famennian).

Understanding the evolution of the apparently very unfern-like plant body of Psilotales, particularly that of Psilotum itself, until fairly recently considered to be perhaps the most primitive extant vascular plant. (When I was taught about Psilotum, it was compared with the Palaeozoic rhyniophytes, now thought to be entirely unrelated.) This still presents difficulties; Kaplan (1997, vol. 3: chap. 10; Siegert 1973 and references) summarize the literature. As Kaplan noted, young leaves of Psilotum did have features of fronds (see also Kaplan 1977). Similarly, the leaves of Equisetum may be secondarily simple. Thus its fossil relative Sphenophyllum has much larger leaves with dichotomous venation; bifacial xylem has been found in this genus (Kenrick & Crane 1997; Doyle 2013 and references).

Ecology & Physiology. The evolutionary physiology of ferns is repaying examination. The tracheids of ferns tend to be long and wide, and the scalariform perforation plates may extend the length of the cell, and as a result, water transport can be relatively efficient (Pittermann et al. 2015); conifers, another group without vessels and that also have quite efficent water transport at least has secondary thickening and tracheids with torus:margo pits (Pittermann et al. 2011). Fern richness in Australia is correlated with precipitation, and depending on what aspect of diversity is emphasized, precipitation may be linked to climate equability, etc., bryophytes, but not flowering plants, showing rather similar patterns (Nagalingum et al. 2015).

Epiphytic ferns are concentrated in Hymenophyllaceae, Polypodiaceae (grammitid ferns), Pteridaceae (e.g. Vittaria) and Dryopteridaceae (e.g. Elaphoglossum) (Schuettpelz & Pryer 2009; Watkins & Cardelús 2012; Kato & Tsutsumi 2013; Rothfels & Schuettpelz 2014; see also above). The gametophytes of these epiphytes are often strap-shaped and long-lived, and, like other fern gametophytes, are dessication-tolerant (Watkins et al. 2007a), the epiphytic species being the most tolerant of drying (see also Nayar & Kaur 1971: survey of gametophyte diversity; Dassler & Farrar 1997, 2001; Farrar et al. 2008; Watkins et al. 2007; Rothfels & Schuettpelz 2014; Farrar 2016). Gemmae on these gametophytes (e.g. Pryer et al. 2016; Farrar 1974, 2016) further increase their longevity. These long-lived strap-shaped or filamentous gametophytes may persist in sites that are hundreds of miles from the nearest sporophytes (e.g. Farrar 1967; Ebihara et al. 2013). (With the advent of the ability to identify gametophytes directly by molecular sequencing, rather than waiting for them to produce sporophytes, this phenomenon is turning out to be remarkably common (Ebihara et al. 2013), and it has interesting implications for the evolution of ferns (see e.g. Ebihara et al. 2009).) Some of these long-lived gametophytes never produce sporophytes, or produce sporophytes only in parts of their ranges (Pinson et al. 2016). Indeed, in places there is a mismatch between altitude and the general diversity of fern sporophytes and gametophytes, sporophytes being most diverse at ca 1,000 m, gametophyte diversity being independent of altitude (Nitta et al. 2017: Polynesia).

Mycorrhizae are less frequent in epiphytic than in terrestrial ferns (e.g. B. Wang & Qiu 2006; Kato & Tsutsumi 2013), and are uncommon in the strap-shaped gametophytes so common in epiphytic ferns (Pressel et al. 2016), and also in the sporophytes of epiphytic Hymenophyllaceae, where dark septate endophytes are commonly to be found (Lehnert & Krug 2019). However, the sporophytes of grammitid ferns have an apparently secondary association with ascomycetes; these ferns tend to be small and are twig epiphytes and they seem to be dependent on this association (Lehnert et al. 2016).

Dessication tolerance is quite common in fern sporophytes, both epiphytic and terrestrial (e.g. Lösch et al. 2007), and peltate scales may play a central role in the uptake of water, rather as in Bromeliaceae (John & Hasenstein 2016). Ferns show a variety of adaptations to the ecologically dry epiphytic habitat, including CAM photosynthesis (Keeley & Rundell 2003; Watkins & Cardelús 2012), and in some respects epiphytic ferns are ecologically more like angiosperms than terrestrial ferns, using water quite efficiently (and having very low hydraulic conductivity) (Watkins & Cardelús 2012), while some ferns, either in the sporophytic or gametophytic stage, show poikilohydric behaviour (e.g. Lösch et al. 2007). For Late Embryogenesis Abundant (LEA) genes and gene families, involved in dessication tolerance in land plants in general, see Artur et al. (2018); little is known about such genes in ferns.

In seed plants, stomata open in response to both red and blue light. Stomata of those few ferns examined, all leptosporangiates (Pteris, Adiantum, Asplenium and Nephrolepis), lacked a blue light-specific opening response, although elements of the response, including the relevant phototropin, a blue right receptor protein kinase, were present (Doi et al. 2006). In the shaded understory environments that many ferns prefer, blue (and red) light has been preferentially absorbed by the canopy above; stomatal closure is mediated by red light. Doi et al. (2015) later found that there was a blue-light response in other ferns, lycophytes, cycads, etc., indeed, stomata of a few of these taxa, including Equisetum, did not respond to red light. Kawai et al. (2003; see also Suetsugu et al. 2005) noted that there was a distinctive chimaeric red/far red light photoreceptor (phy 3, = neochrome) in which red-sensing phytochrome and blue-sensing phototropin are fused into a single molecule (F.-W. Li et al. 2014) in some Polypodiales, possibly aiding in phototropic responses in the shaded conditions in which ferns have diversified; neochrome is very sensitive to white light (Wada 2008). Neochrome seems to have been acquired by polypod ferns from hornworts via lateral gene transfer around 179 Ma, which, perhaps not so coincidentally, is about the age of Polypodiales (H. Schneider et al. 2004a), and it may even have been acquired more than once (F.-W. Li et al. 2014); it is also known from Mougeotia, Zygnemataceae (Suetsugu et al. 2005). Veins in fern leaves are not the same distance from the epidermis as they are from each other, i.e., they are not hydraulically optimised. However, photosynthesis may proceed best in the low levels of diffuse light in understory vegetation and so this hydraulic optimisation may not matter so much (Zwieniecki & Boyce 2014). (Note that in seed plants the effectiveness of the stomatal response is white/blue + red > blue > red > green light (Willmer & Fricker 1996) and in many flowering plants leaves are hydraulically optimised - Zwieniecki & Boyce 2014.) For iridescence in fern leaves, and different kinds of chloroplasts in epidermal cells, see R. M. Graham et al. (1993) and Nasrulhaq-Boyce et al. (1991); being shade plants, they may have distinctively coloured/patterned fronds, show iridescence, etc..

Stomatal responses in non-flowering plants have been characterized as being passive, those of flowering plants, active. Thus it has been suggested that fern stomata do not respond to increased CO2 concentrations in the atmosphere (McAdam & Brodribb 2012a), although this has been questioned, such a response apparently being present in the common ancestor of extant vascular plants (Franks & Britton-Harper 2016). Furthermore, the synthesis of abscisic acid, involved in stomatal responses to CO2 concentrations and other stimuli, is not triggered by changes in leaf turgor in non-flowering plants, but it is in flowering plants, although not all of the latter show the same sensitivity to turgor changes (McAdam & Brodribb 2016). Interestingly, the stomata of Equisetum seem to be permanently closed, as perhaps also in Psilotum (Cullen & Rudall 2016). However, details of the physiology of stomatal functioning is unclear and in part conflicting (Lind et al. 2015; Roelfsema & Hedrich 2016 and references).

Campany et al. (2018) compared ferns (6 species) with Selaginella (7 species) growing in similar habitats on the forest floor in Costa Rica, and found that the two groups had similar net photosynthesis and water use efficiency, but the latter had a lower leaf mass/area and nitrogen use efficiency in photosynthesis, and longer, less dense stomata.

Around 38-43% of all ferns, epecially members of the [Marattiidae + Polypodiidae] clade, accumulate aluminium (i.e. tissue concentration >1000 mg kg-1), a far higher percentage than in seed plants (Chenery 1949a, b; Schmitt et al. 2017).

Fertilization & Spore Dispersal. The sex of gametophytes may be controlled by pheromones (Atallah & Banks 2015).

Noblin et al. (2012) describe the mechanics of how the annulus functions in sporangium dehiscence.

Plant-Animal Interactions. Herbivory in ferns is much lower compared with angiosperms, and only some 35% of the herbivores are restricted to ferns (Hendrix 1980; Ehrlich & Raven 1964; c.f. Balick et al. 1978: herbivory about the same, 73% restricted; Turcotte et al. 2014: caveats). Ferns quite commonly have phytoecdysones, which can defend against many herbivores (Hikino et al. 1973; Balasubramanian et al. 2006), or their may be fairly specifically-targeted protectants (Shukla et al. 2016). Interestingly, in the chemically well-defended bracken, Pteridium aquilinum, herbivory is not associated with jasmonate-mediated regulation of VOC (volatile organic compound) emissions, which in flowering plants, for instance, can attract predatores of herbivores, etc. (Radhika et al. 2012). For herbivory by caterpillars of lithinine geometrid moths, which has involved host shifts after a single colonization event, see Weintraub et al. (1995).

Bacterial/Fungal Associations. For mycorrhizae in ferns, see Lehnert et al. (2010, 2016: "endophytes" s. l. include mycorrhizae and endophytes s. str.) and Pressel et al. (2010, 2016). Fine endophytes, with their distinctive fan-like arbuscules, are known from ferns; these endophytes are probably mucoromycotina, not Glomus (Orchard et al. 2016 and references). Mycorrhizal associations are commoner in ferns than in lycophytes (Pressel et al. 2010; Rimington et al. 2014: ca 2/3 vs <1/2 species examined; see also Lehnert et al. 2016). Mycorrhizal fungi may be found in heat-shaped fern gametophytes where they gain entry via the rhizoids and live in the central cushion, but they are not found in taxa with filamentous or strap-shaped gametophytes (e.g. Pressel et al. 2010, see above). For glomeromycote associations with fern gametophytes, see Pressel et al. (2016); note that relatively little is known about such associations in Polypodiales. Still less seems to be known about associations of fungi with the sporophytic generation, but in those with very thiny, wiry roots ca 1 mm across mycorrhizae are uncommon, fungi with dark septate hyphae have been reported, especially in epiphytic taxa (Lehnert et al. 2916), but Pressel et al. (2010, 2016) suggests that they are unlikely to be members of mutualistic associations. Equisetum is not known to form any mycorrhizal associations (Read et al. 2000; Pressel et al. 2010, 2016, c.f. Lehnert et al. 2016), probably because it is frequently to be found in more or less transiently wet but nutrient-rich habitats (see also Salviniales, liverworts in similar habitats).

Vegetative Variation. Vasco et al. (2013) summarize fern frond morphology and development, noting a number of shoot-like features. Details of leaf development in Equisetum are similar to those of other ferns and seed plants rather than of lycophytes (Ambrose & Vasco 2016), but it has paracytic stomata of a kind unlike those of any other land plants (Cullen & Rudall 2016).

Genes & Genomes. Ferns are noted for their high incidence of polyploidy, almost 1/3 (31%) of all speciation events being accompanied by polyploidy (Wood et al. 2009). Thus there has been recent hybridization between clades (Cystopteris, Gymnocarpium) that diverged an estimated (76.2-)57.9(-40.2) Ma (Rothfels et al. 2015a) while hybridization in Osmunda has occurred between clades that may have been separated for four times as long (Bomfleur et al. 2014b; Grimm et al. 2015).

For information on genome size, see Obermayer et al. (2002), J. Clark et al. (2016: many details), Hidalgo et al. (2017c) and Pellicer et al. (2018: mean 1C genome = 14,320 Mb); particularly large genomes occur in Psilotales (Hidalgo et al. 2017b) and some Ophioglossales, and in a few polypods. Clark et al. (2016) suggested that the epiphytic habit and increased genome size in eupolypods may be connected, but this would be the reverse of the situation in orchids. Polyploidy is common here, but the amount of DNA per chromosome tends to be conserved, unlike in angiosperms (see also Pellicer et al. 2018). Heterosporous ferns tend to have lower chromosome numbers and lower genome sizes that homosporous ferns, as in other vascular plants (Sessa & Der 2016 and references; F.-W. Li et al. 2018). Genome size evolution was found to be higher than expected, being higher even than that in several angiosperm groups - for example, higher than in Brassicales, lower than in Caryophyllales (Baniaga et al. 2016), although earlier work had found relative stasis (e.g. Clark et al. 2016).

For the evolution of the monilophyte chloroplast genome, see Karol et al. (2010), Grewe et al. (2013), and Kuo et al. (2018b). Robison et al. (2018) emphasized the importance of mobi;le open reading frames in affecting variation in the chromosome genome of Pteridaceae is particular, but these are more widesly distributed in ferns (and elsewhere) and may also be found in mitochondria and nuclei. Some inversions in the chloroplast inverted repeat and other changes in the chloroplast genome may be high-level synapomorphies (Gao et al. 2009; Ruíz-Ruano et al. 2018). Grammitidaceae s. str. in particular (included in Polypodiaceae below) have green spores and accelerated plastid genome evolution, a correlation found elsewhere in ferns, although it is not 100% (H. Schneider et al. 2004b), indeed, species with spores that less obviously contain chloroplasts are quite widespread (Sundue et al. 2011).

Chemistry, Morphology, etc.. The xyloglucan composition of the primary cell wall varies substantially, that of Equisetum being particularly distinctive (Hsieh & Harris 2012); for cell wall polysaccharides, see also Silva et al. (2011).

Basic sporophyte morphology was outlined by Kaplan (1997, vol. 2: chap. 11, 18, vol. 3: chap. 19, 2001).The information on horizontal cell walls in early embryo/sporophyte development in ferns given by Philipson (1990) seems to be incorrect - the examples should be vertical? For the organization of the apical meristem of the stem, see Ambrose and Vasco (2015), also Tracheophytes above. The stem has a siphonostele, the protoxylem being restricted to lobes of the central xylem strand, hence bringing to mind a necklace (development of the xylem is mesarch, although it is notably variable in the Ophioglossum/Psilotum clade). The protoxylem is described as having G-type tracheids (Edwards 1993). Hernández-Hernández et al. (2012) discuss the distribution of the circumendodermal band, tannin-containing cells with thickenings on in inner periclinal and also sometimes the anticlinal walls, that surround the petiolar vascular bundles and that have a common orgin with the endodermis; they also detail the distributions of a number of other vegetative/habit features. For aerophore distribution in ferns, see Davies (1991), while Wagner (1979) discusses reticulate venation in fern fronds - a small amount of reticulation is common, but not extensive reticulation.

Takahashi et al. (2009, 2014 and references) describe gametophyte development in ferns. They note that the apical region converts to a multicellular meristem, which can divide - dichotomous branching - if cell division in the middle of the meristem stops; the branched, strap-shaped gametophytes of epiphytic ferns are simply an extreme variant of this morphology. Archegonia develop only after the formation of the multicellular meristem. For details of male gamete morphology and movement, etc., see e.g. Renzaglia et al. (2000b, 2002) and H. Schneider et al. (2002).

For information on pteridophytes in general ("pteridophytes" in the past have often - and still may - include lycophytes), see also Kato (2005) and Ranker and Haufler (2008). For other general information, see Raven and Edwards (2001), for comparative anatomy, see Ogura (1972), for vessels, see Sen and Mukhopadhyay (2014 and references), for phloem, see Evert (1990a), for venation, see Wagner (1979) and Boyce (2005b), for details of stelar morphology and evolution, see Beck et al. (1982), and for young sporophytes, etc., see Johnson and Renzaglia (2009 and references).

Phylogeny. The overall circumscription of the fern clade has only recently become clear, even if the actual position of Equisetum is somewhat uncertain (the three hypotheses below) and a number of details of family relationships need clarification.

1. Equisetum may be embedded in ferns, perhaps sister to Angiopteris, etc. (although support only moderate), the combined clade in turn being sister to remaining ferns (e.g. Pryer et el. 2001a, 2004a; Wikström & Pryer 2005; Qiu et al. 2007; Ebihara et al. 2011; Evkaikina et al. 2017; c.f. in part Wolf et al. 1998). Lehtonen (2011: some analyses) and Z.-D. Chen et al. (2016) found that it was sister to the eusporangiate ferns, although support was not strong, indeed, the latter group recovered Marattiales as sister to all other ferns Hennequin et al. (2008) recovered relationships [[Ophioglossales + Psilotales] [Marattiales [Equisetales ...]]], while in some analyses Lehtonen (2011) found a position sister to [Marattiales + eusporangiate ferns].

2. Equisetum may be sister to all other ferns, as in a rps4 analysis, and also 4- and 5-gene analyses, the latter two with strong support (Schuettpelz et al. 2006), analyses of several plastid genes (Rai & Graham 2010), and in a matK phylogeny (Kuo et al. 2011). Knie et al. (2015) also find good support for the relationships [Equisetum [[Psilotum + Ophioglossum] + The Rest]], as does Rothfels et al. (2015b) in their nuclear gene analysis and Testo and Sundue (2016) in their 4,000 species analysis; see also Labiak and Karol (2017: most analyses). This sister group relationship of Equisetum was also shown by Sessa and Der (2016), Qi et al. (2018) and references) and H. Shen et al. (2017), the last two studies using transcriptome data, support was strong. Indeed, this relationship is quite commonly found. Interestingly, spore wall ultrastructure of Calamites, an extinct equisetaceous plant, is not so different from that of Ophioglossaceae and other ferns (Lugardon & Brousmiche-Delcambre 1994; Grauvogel-Stamm & Lugardon 2009). Equisetum has no mitochondrial atp1 intron, either a secondary (and parallel) loss or plesiomorphic absence, depending on the topology of the whole group (Wikström and Pryer 2005: see the character hierarchy below).

3. H. Schneider et al. (2009) noted potential morphological apomorphies such as simple leaf blade and stems with both radial and dorsiventral symmetries (= erect plus creeping stems...) suggesting a clade [Psilotales + Equisetales], consistent with some structural changes in the chloroplast genome (Grewe et al. 2013; see also some analyses in Karol et al. 2010; also Wolf & Karol 2012; Ruhfel et al. 2014; J.-M. Lu et al. 2015; Gitzendanner et al. 2018 and Lehtonen et al. 2018, both chloroplast genomes; references in Qi et al. 2018; Kuo et al. 2018b: plastomes).

Below Equisetum alone is placed as sister to all other monilophytes (hypothesis 2), although hypothesis 3 also has support. The strongly supported [Psilotum + Ophioglossum] clade (Tmesipteris is sister to Psilotum) is perhaps sister to all other ferns, as chloroplast data has broadly tended to suggest (Rothfels et al. 2015b for references). The use of morphology alone or in combination with molecular data affects the relationships detected (Wikström & Pryer 2005 and references); see also Grand et al. (2013) for various morphological analyses. For additional details of relationships in ferns, see the discussion below.

This reorganisation of monilophytes has sometimes been severely criticised (Rothwell & Nixon 2006), but it is unclear how damning such criticism is. Since the evaluation of "support" values for particular topologies is integral to the approach adopted in these pages, the decision to exclude such values by those authors makes their work difficult (for me, at least) to interpret. Indeed, in several morphological cladistic analyses (e.g. Bremer 1985; Stevenson & Loconte 1996; Rothwell 1999: fossils included or not) Psilotum came out as sister to all other vascular plants, and although some morphological analyses (e.g. H. Schneider et al. 2009) do place Psilotum with other monilophytes, the same analyses also place flowering plants within a paraphyletic group of extant gymnosperms... You win some, lose some.

Other than attempts to clarify these basal relationships, there has been a fair bit of work on other relationships in the fern clade. Qi et al. (2018) recently carried out a series of analyses on 136 genera (in 43 families) using a variety of data sets, the smallest having 72 nuclear genes and the largest having 935 genes. Their results are largely consistent with those described below, differences being noted where they are important. The trees produced by H. Shen et al. (2017) in a phylogenomic analysis of 69 taxa also differed somewhat. Thus Shen et al. (2017; see also Wickett et al. 2014) obtained a moderately well supported [Marattiales + Psilotales] clade sister to leptosporangiate ferns, but this may be a sampling issue. See also Qi et al. (2018) and Kuo et al. (2018b: plastomes), in the latter, relationships are broadly similar to those below, except in the position of Equisetm (see above) and in that of Pteridium (sister to the Eupolypod II clade), but again sampling is skimpy.

Classification. A. R. Smith et al. (2006, 2008) propose a phylogeny-based reclassification of the ferns, and they also include literature, ordinal and familial synonymy, and a list of accepted genera and some major synonyms; Prelli (2010) gives a nice account of European ferns. However, adjustments to this classification are being made as details of the phylogeny become better understood (Schuettpelz & Pryer 2007, 2008; Kuo et al. 2011; Rothfels et al. 2012b: reclassification of eupolypods II; Rothfels et al. 2015b). Here I follow The Pteridophyte Phylogeny Group (2016) which should be consulted for details, as the authors make it clear that the monophyly of some genera they accept is unclear - or the genera may even be polyphyletic; further instalments of this classification are eagerly awaited. A linear sequence of families and genera (Christenhusz et al. 2011a) is now dated, but more recently Christenhusz and Chase (2014) have proposed another classification as has H.-M. Liu (2016). There are substantial differences between these classifications, and this has occasioned some controversy - c.f. Schuettpelz et al. (2018) and Christenhusz and Chase (2018), splitting vs lumping.

Previous Relationships. Psilotum and Equisetum had earlier been thought to represent independent lineages, with Psilotum and relatives considered to be the most primitive living vascular plants, and the latter do look superficially similar to some early fossils. Their association with ferns, now very largely accepted, was unexpected (but see Kenrick & Crane 1997). Although Bierhorst (1968, see also 1977) had compared Psilotum with the extant fern Stromatopteris (Gleicheniaceae) and found some morphological similarities, most of these have turned out to be parallelisms and the two are not at all closely related.


EQUISETOPSIDA / [Equisetidae + Ophioglossidae] (assuming this clade exists): plant with erect and creeping stems; ?leaf vernation; tapetum plasmodial; embryo exoscopic, suspensor 0; chloroplast rps16 gene and rps12i346 intron lost.

Age. The clade that contains Equisetum has probably been separate from other monilophytes since the Permian, ca 250+ Ma (Stanich et al. 2009); B. Zhong et al. (2014b) thought that this clade was (370.3-)296.2(-189.9) Ma old.

EQUISETIDAE Warming

EQUISETALES Tavis.

Just the one family, 1 genus, 18 species.

EQUISETACEAE L. C. Richard

Plant with erect and creeping stems; roots not mycorrhizal, triarch, with large central tracheid; cell walls also with (1→3),(1→4)-ß-D-MLGs [Mixed-Linkage Glucans], SiO2 accumulation common; stem with intercalary meristem [at base of leaf sheath], ridged, photosynthetic, protoxylem mesarch, with central canal; protoxylem lacunae developing; stomata paracytic, maturing basipetally, mesogenous, subsidiary cells forming stomium; leaf vascular bundles amphicribral; leaves whorled, "axillary" buds +, alternating with the leaves, members of branch whorls alternating at each node; leaves small, simple, 1-veined, basally connate, not photosynthetic; sporangiophores peltate, aggregated into a strobilus; sporangial cell walls with helical secondary thickenings; tapetum plasmodial; spores with circular aperture [hilate], abapertural obturator +, green, wall with silica, elaters 4-6/spore, spatulate, helically-coiled; gametophyte not mycorrhizal, (bisexual), unisexual (heteromorphic); embryo exoscopic, plane of first cell division variable, suspensor 0; n = 108; mitochondrial atp1 intron 0.

1 [list]/18. Northern Hemisphere, southeast Asia and Malesia to New Caledonia, South America, the Canaries, East African mountains and the Mascarenes.

Age. Extant species of Equisetum seem to have separated in the Caenozoic (77.5-)64.8(-52.1) Ma (Des Marais et al. 2003, but c.f. Stanich et al. 2009); Elgorriaga et al. (2018 and references) suggest that divergence might have begun some time between the Early Jurassic and Early Cretaceous 190-140 Ma.

Fossils with many of the apomorphies of crown group Equisetum are known from Upper Jurassic deposits from Patagonia some 150 Ma or more old (Channing et al. 2011; see also Stanich et al. 2009); for still older Equisetum-like spores - but with trilete marks - and associated elaters from the Middle Triassic, see Schwendemann et al. (2010).

Evolution: Divergence & Distribution. Change in spore morphology from the Calamites type to the at first sight very different trilete spores of Equisetum is convincingly demonstrated by Grauvogel-Stamm and Lugardon (2009). Other fossils placed in this area include Sphenophyllum, which has megaphylls, and this suggests that the tiny leaves of Equisetum may be reduced rather than being microphylls; roots of fossils also differ in morphology and can be polyarch. For the long fossil record of Equisetum, see Channing et al. (2011) and Elgorriaga et al. (2018), the latter including a number of extinct Sphenopsida in their trees.

Infrageneric groups and possible apomorphies are discussed by Elgorriaga et al. (2018).

Ecology & Physiology. Equisetum tends to grow in ecologically rather stressfull habitats, including hot springs (Channing et al. 2011 and literature; Husby 2012). There is pressurized gas flow, so-called convective ventilation, from the stems into the rhizomes in some extant species of Equisetum and probably in their fossil calamitalean relatives from the Carboniferous. Oxygen moves via interconnected air spaces into the rhizomes - and this is accompanied by changes in position of the flow-resistant endodermis in the stem - so perhaps allowing them to penetrate deeply into the anoxic substrates commonly favoured by this group (Armstrong & Armstrong 2009). Species of Equisetum that lack interconnected air spaces have no convective ventilation, yet these species like E. hyemale may still grow in anoxic, partly submerged conditions (Armstrong & Armstrong 2011). Indeed, there is a literature suggesting that the complex system of air canals is involved in plant support (Spatz et al. 1998; Speck et al. 1998).

The stomata are likely to be immobile, especially when older, because of the rigid radiating rib-like silica thickenings of the subsidiary cells, and the stomata may even be permanently closed then, although the chloroplasts can move from one guard cell to another (Cullen & Rudall 2016). For silica deposition, see also Law and Exley (2011); note that stomata - and the plant (Equisetum arvense) - seem to be able to develop without silica.

Reproductive Biology. Aided by their elaters, spores of Equisetum can either jump up to 1 cm in the air as they dry, or move by short random walk-type movements along the ground (Marmottant et al. 2013). Studies on the stem of E. hyemale focussing on anatomy, node number, etc., suggested that how the stem vibrates may affect spore dispersal (Zajaczkowska et al. 2017).

Bacterial/Fungal Associations. For the absence of mycorrhizae here, see Pressel et al. (2016, c.f. Lehnert et al. 2016).

Genes & Genomes. For a genome duplication in Equisetum dated to (112.5-)92.4(-75.2) - or perhaps only 70-50 - Ma, see Vanneste et al. (2015).

Chemistry, Morphology, etc.. For the absence of mycorrhizae, see Pressel et al. (2016). For (1→3),(1→4)-ß-d-glucans, see Fry et al. (2008) and Xue and Fry (2012), and for Si concentration, etc., see Husby (2012). Mixed-linkage glucans are known from Equisetum (Fry et al. 2008; Xue & Fry 2012). For leaf morphology, see Rutishauser and Sattler (1987). Cullen and Rudall (2016) discuss the development of the remarkable stomata in detail; they are mesogenous. Interestingly, E. arvense, at least, can grow apparently healthily in the total absence of silicic acid (Law & Exley 2011). Sporophyte growth is discussed by Tomescu et al. (2017), for the gametophyte, see Hauke (1969).

The tapetum seems to be involved in the formation of the elaters (Uehara & Murakami 1995).

Phylogeny. Equisetum bogotense is sister to the rest of the genus (Guillon 2007; Christenhusz et al. 2019), and a position as a distant sister to the rest of the genus was recovered in molecular and joint, but not morphology-only, analyses that were carried out by Elgorriaga et al. (2018).

Classification. The two-subgenus classification of Equisetum (e.g. Hauke 1978) will need amending given the position of Equisetum bogotense as sister to the rest of the genus, and Christenhusz et al. (2019) do the job, however, previous sectional classifications have not held up.

[[Ophioglossidae] [Marattiidae + Polypodiidae]]: sporophyte (with glomeromycote associate); sporangia not aggregated into strobili, (spore aperture proximal and monolete); sporophyte-gametophyte boundary with unicellular sporophytic haustoria [?level]; (gametophyte with glomeromycote associate); indel in mitochondrial rpl2 coding region.

Age. Rothfels et al. (2015b) suggested an age of (390.4-)351.7(-309.7) Ma for this node.

OPHIOGLOSSIDAE Klinge

Plant with erect and creeping stems; stem protoxylem development variable; embryo exoscopic, suspensor 0; gametophyte subterranean, axial, mycoheterotrophic [non-photosynthetic, mycorrhizal].

Age. Magallón et al. (2013) estimated an age of around 275.6 Ma for this clade, and (316-)306(-267) Ma is the age in Y. L. Qiu et al. (2007), ca 214.5 Ma in Laenen et al. (2014) and (317.1-)250.5(-141.8) Ma in Rothfels et al. (2015b); see also P. Soltis et al. 2002) for estimates.

Evolution: Bacterial/Fungal Associations. For mycoheterotrophy in this clade, not yet confirmed for the small genera in Ophioglossaceae, see Merckx et al. (2013a) and Pressel et al. (2016).

Genes & Genomes. Both Psilotum and Ophioglossum have large genomes (Bennett & Leitch 2005; J. Clark et al. 2016).

For the mitochondrial genomes of Psilotum (very large) and Ophioglossum, see W. Guo et al. (2016b); they are similar in terms of apomorphies to those of seed plants.

PSILOTALES Prantl

Just the one family, 2 genera, 17 species.

PSILOTACEAE J. W. Griffith & Henfrey

Epiphytes; silica content slight; roots 0; ?leaf vascular bundles; leaves small, simple, (laterally flattened - Tmesipteris), veins 1 or 0; sporangia 2-3, forming synangium; tapetum glandular-amoeboid; spores kidney-shaped, monolete; gametophyte with septate rhizoids; (transfer cells in sporophyte only - Psilotum), n = 52, 208, nuclear genome [1C] 72.5-150.6(?-313) pg/147.3 Gb.

2 [list]/17. Pantropical, inc. Pacific, warm temperate.

Age. B. Zhong et al. (2014b) estimated an age of (147.1-)72.3(-14.7) Ma and Rothfels et al. (2015b) an age of (142.5-)78.9(-28.5) Ma for this clade.

Evolution: Bacterial/Fungal Associations. Mycorrhizal associations in Psilotum in the echlorophyllous gametophytic and subterranean sporophytic stages are with glomalean Glomus group A fungi, and they are also to be found in the autotrophic sporophyte (Winther & Friedman 2009; Imhof et al. 2013 and references).

Genes & Genomes. For the huge genome of Tmesipteris obliqua, see Hidalgo et al. (2017b, c).

OPHIOGLOSSSALES Link

Just the one family, 4-10 genera, 125 species.

OPHIOGLOSSACEAE Martinov

Roots fleshy ca 2 mm< across, with 2-5 protoxylem poles; root hairs 0; cork mid cortical; stem stele sympodial; tracheids with circular bordered pits; frond vascular bundles collateral; (axillary buds +); fronds compound to simple, 1 produced/year, petiole vasculature U-shaped, venation reticulate, with internally directed veins, vernation nodding, bases sheathing, stipules +, thin; one or more sporophores associated with each tropophore; (gametophyte with septate rhizoids); (embryo endoscopic; first cell wall of the zygote vertical); n = (44) 45 (46 ...720), nuclear genome [1C] ca 28.35(?-139) pg.

4-10 [list]/125: Botrychium (42). More or less world-wide.

Age. Rothfels et al. (2015b) suggested an age of (249.6-)161.7(-74) Ma for crown-group Ophioglossaceae, Gil and Kim (2018) an age of ca 256 m. years.

Evolution: Divergence & Distribution. Mankyua, ca 194.8 Ma, is known only from Jejudo Island ca 2 Ma - although in contact with the Korean mainland during its existence (Gil & Kim 2018). There are two bipolar disjunctions within Botrychium in which polyploidy/introgressive hybridization has been involved; such distributions are uncommon in ferns, but interestingly, the gametophytes can self (Farrar & Stensvold 2017; see also Dauphin et al. 2017a). Species tend to be donors of either paternal or maternal genomes (Dauphin et al. 2017b, esp. Fig. 4).

Ecology & Physiology. Glomeromycote mycorrhizae in Ophioglossum and Botrychium are associated with the echlorophyllous gametophyte and subterranean sporophytic stage, and also the photosynthesising sporophyte; the latter may obtain some of its carbon from other conspecific or heterospecific plants with which the fungus is associated (Winther & Friedman 2007; Field et al. 2015a; Field & Pressel 2018). Botrychium gametophytes do not develop beyond the 8-celled phase without establishing this mycorrhizal association, and plants may spend the first ten years of their lives underground (Winther & Friedman 2007), and even when above-ground shoots are produced, these may not appear every year (Reintal et al. 2010; Shefferson et al. 2018).

Bacterial/Fungal Associations. The Glomus involved in the mycoheterotrophic association is in the same immediate clade as species involved in similar associations in Arachnitis (Corsiaceae) and Lycopodiaceae, and the fungi are the same as those associatd with the autotrophic sporophyte (Winther & Friedman 2007, 2008; Imhof et al. 2013).

Genes & Genomes. Mitochondrial genes from a presumably root-parasitic member of Loranthaceae seem to have been acquired by the fern Botrychium virginianum, perhaps via a common mycorrhizal associate (Davis et al. 2005b). However, in general mycorrhizae are not very common on Santalales and if the gene transfer took place in Asia where the fern may have originated, as Davis et al. (2005b) suggested, the absence of extant root-parasitic Loranthaceae from that area is notable.

Ophioglossum reticulatum, at n = 720, has the highest chromosome numbers of any plant.

Chemistry, Morphology, etc.. Takahashi and Kato (1988) describe the development of lateral meristems in the family. There is some debate as to whether there is secondary thickening in Botrychium. It certainly is not conventional secondary thickening, and the distinctive appearance of the vasculature tissue - the tracheary elements are in radial series, even if there is no evidence that the stem has increased in width - may be connected with the fact that the leaves may take up to five years to develop, perhaps a record for land plants (Rothwell & Karrfalt 2008).

Phylogeny. See Hauk et al. (2003) for a phylogeny, Mankyua not included, also Shinohara et al. (2013), Mankyua included, but position unstable - sister to rest of family (also in the joint analysis), or to Ophioglossum s. str.. Mankyua continues to present problems - it was sister to the rest of the family in maximum parsimony analyses, while Psilotum took that position in other analyses (Gil & Kim 2018).

Classification. Somewhat in a state of flux - Mabberley (2008) noted there were four genera and 55 species, as of iv.2017 I had 4 genera and 80 species, while The Pteridophyte Phylogeny Group (2016) recognized 10 genera with 112 species.

[Marattiidae + Polypodiidae]: plants often Al accumulators; frond vascular bundles amphicribral; frond compound, vernation circinate; scales +; sporangia abaxial; gametophyte green, surficial; shoot developing towards the archegonial neck [from hypobasal cell, endoscopic]; nuclear genome 3.5-14.0 pg; changed gene adjacencies at borders of chloroplast IR; mitochondrial atp1i361g2 intron gain.

Age. B. Zhong et al. (2014b) suggested an age of (378.8-)336.7(-291.5) Ma and Rothfels et al. (2015b) an age of (364.1-)329(-289.2) Ma for this clade.

P. Soltis et al. (2002) offer a variety of suggestions for ages of nodes in this clade.

MARATTIIDAE Klinge

Roots with several protoxylem poles; root hairs septate [?multicellular]; petiole vasculature polycyclic; stipules +, fleshy and starchy.

MARATTIALES Link

Synonymy: Christenseniales Doweld

Just the one family, 6 genera, 110 species.

MARATTIACEAE Kaulfuss

Roots fleshy, ca 2 mm≤ across, root hairs few; mycorrhizae +; dictyostele +; mucilage canals +; rhizome with scales; aerophores linear, with lenticels; petiole vasculature polycyclic; fronds pulvinate, (divisions with internally directed reticulate venation - Christensenia); meiosis monoplastidic [?all]; spores bilateral or ellipsoid, monolete; transfer cells 0; x = 40.

6 [list]/110: Danaea (50).

For meiosis, see Brown and Lemmon (2001).

Evolution: Divergence & Distribution. For some comments on biogeography, see Christenhusz and Chase (2013). There seems to be a slow-down in the rate of evolution in this clade (Rothfels et al. 2015b and references).

Ecology & Physiology. Aluminium accumulation is quite common in Marattiaceae (Schmitt et al. 2017).

Chemistry, Morphology, etc.. For gum in Marattia, see Lambert et al. (2016). H. Shen et al. (2017) scored both Osmunda and Angiopteris as having a rudimentary annulus.

Phylogeny. For a phylogeny, see Murdock (2008a), also Christenhusz et al. (2008); the fossil Psaronius associates consistently with Marattia (e.g. Grand et al. 2013 and references). Rothwell et al. (2018b) include fossils in their comprehensive study of the family, but relationships between the extant genera become unclear depending on data/fossils included and the outgroup used; Psaronius and several other fossils are in a clade sister to the clade that includes all extant Marattiales along with a few fossils.

Both Marattia and Angiopteris are paraphyletic, but they can easily be made monophyletic (Murdock 2008b). For Danaea, see Christenhusz (2010).

POLYPODIIDAE Cronquist, Takhtajan & Zimmermann / leptosporangiate ferns.

Blue light stomatal opening response absent; primary cell walls poor in mannans and rich in tannins; roots with 2 protoxylem poles; primary xylem with scalariform bordered pits; leaf trace single; aerophores linear, on either side of the petiole, with stomata; sporangium derived from periclinal division of a single epidermal cell, wall one-layered, stalk 4-6 cells across [= leptosporangium]; sporangium with exothecium forming an annulus, 64-800 spores/sporangium; antheridium ± exposed; gametophyte cordate [level?]; embryo prone [first cell wall of the zygote vertical, parallel to gravity], with quadrant/octant formation, suspensor 0.

Age. Magallón et al. (2013: with temporal constraints) estimated an age of around (267.8-)252.7-251.4(-246.1) Ma for this clade; ca 299 Mais the age in Schuettpelz and Pryer (2009), (327.8-)301.3(-271.5) Ma in Rothfels et al. (2015b), ca 309.5 Ma in Hennequin et al. (2008), (330-)323(-310) Ma in Y. L. Qiu et al. (2007), perhaps 350 Ma in H. Schneider et al. (2004a), and (357.5-)357(-356) Ma in Testo and Sundue (2016) - but only around 170 Ma in Villarreal and Renner (2014). All told, a rather disconcerting spread.

Ecology & Physiology. Tolerance of extreme dessication, sometimes facultative, is scattered through this clade, and this occurs in gametophytes, too (Proctor & Tuba 2002; Gaff & Oliver 2013).

Genes & Genomes. The GC content of the chloroplast genome is notably high around here (Y. Yu et al. 2019b).

Phylogeny. The large clade made up of leptosporangiate ferns has very strong support (see also Hasebe et al. 1994, 1995, Pryer et al. 1995; Wolf et al. 1998; Quandt et al. 2004; Schuettpelz et al. 2006; Rai & Graham 2010, etc.). Rothfels et al. (2015b) emphasized that their analyses of nuclear data broadly agreed with several plastid sequence analyses.

Within the leptosporangiates, Osmunda and relatives, the sporangia of which have some eusporangiate features, are strongly supported as being sister to the rest. There is further substantial resolution of relationships (e.g. Pryer et al. 2004a, b and references; Schuettpelz & Pryer 2007; c.f. in part Kuo et al. 2011: positions of Gleicheniaceae, Lindsaeaceae, Nephrolepis [previously in Lomariopsidaceae] uncertain). It is unclear whether or not there is a [Hymenophyllales + Gleicheniales] clade (Knie et al. 2015 for literature; H. Shen et al. 2017; Kuo et al. 2018b; Lehtonen 2018); relationships here are probably best represented as a tritomy. There is some evidence that Gleicheniales are paraphyletic, but support is slight (Rothfels et al. 2015b; see also Qi et al. 2018). J.-M. Lu et al. (2015: chloroplast genome, but sampling) placed Dipteridaceae and Lygodiaceae as successive branches along the leptosporangiate stem. Schizaeales and Salviniales (strong support) and Cyatheales (weak support) are successively sister to Polypodiales (e.g. Rothfels et al. 2015), and Rai and Graham (2010) suggest additional variants. For relationships in Schizaeales, see Labiak et al. (2015 and references). For discussion of relationships within Polypodiales, see below.

OSMUNDALES Link

Just the one family, 6 genera, 18 species.

OSMUNDACEAE Martinov

Roots (≥1 mm across); cataphylls [= petiole bases] +; SiO2 accumulation common; stem with ectophloic siphonostele, with a ring of conduplicate/twice conduplicate discrete bundles; leaves with stipules; petiole vasculature complex U-shaped; fronds with fertile and sterile portions (fertile and sterile fronds separate); annulus a lateral group of cells; spores green; (rhizoids septate), zygote elongating; n = 22.

6 [list]/18. Almost worldwide, not Middle East-Siberia or polar.

Age. The age of this clade is estimated at around 199.6 Ma (Schuettpelz & Pryer 2009) or ca 206 Ma (Hennequin et al. 2008); however, estimates from Carvalho et al. (2013) based on fossils that can be assigned to the leptosporangioid branch of the tree suggest an age in excess of ca 265 Ma, Late Permian (see also Wilf & Escapa 2014), while the preferred age in Grimm et al. (2015: comprehensive analysis, also incorporating fossils) is (264-)243(-233) Ma (see also Bomfleur et al. 2017). It has also been suggested that the Osmunda clade originated in the late Carboniferous, ca 323 or 305 Ma (Phipps et al. 1998; H. Schneider et al. 2004a).

Evolution: Divergence & Distribution. Osmundaceae are very common and diverse in the fossil record from the Permian onwards, but perhaps less so more recently (Grimm et al. 2015). These fossils often have remarkably good preservation, thus a fossil some 180 Ma has anatomy that is remarkably like that of the extant Osmunda claytoniana (Bomfleur et al. 2014a). Detailed studies of axis anatomy have been carried out (Miller 1971; Bomfleur et al. 2017) and used to reconstruct the phylogeny of the group (Bomfleur et al. 2017: some states arbitrary). Indeed, the chromosome number and genome size of 180 Ma fossils similar to Osmunda claytoniana and those of the extant plant may have been similar (Bomfleur et al. 2014a).

H. Schneider et al. (2015) reconstruct the history of genome size changes in Osmundaceae; there is some variation. Generation times in royal ferns are long, so genome change might be expected to be low, but whether or not there is "genomic stasis" (Bomfleur et al. 2014a; c.f. Schneider et al. 2015) is another issue.

Todea, known only from New Guinea, Australia, New Zealand and South Africa, has been found in the early Eocene of Patagonia in rocks ca 53 Ma (Carvalho et al. 2013); see also Fagaceae.

Bomfleur et al. (2014b) note that Osmundastrum cinnamomea is able to hybridize with some species of Osmunda, but not with other Osmundaceae. Given an estimated date of the split of the first two of (264-)238(-233) Ma, only a little less than the age of crown-group Osmundaceae as a whole (Grimm et al. 2015), they have been separated for far longer than any other vascular plants that hybridize.

It has been suggested that Osmunda is paraphyletic, with Osmunda (now = Osmundastrum) cinnamomea being sister to the rest of the family (Metzgar et al. 2008); relationships in Carvalho et al. (2013) are [Osmundastrum [Osmunda [Leptopteris, Todea, Todites]]]. However, Bomfleur et al. (2014b) argue for the monophyly of [Osmunda + Osmundastrum] based on extensive data from fossils and a re-evaluation of the molecular evidence.

The classification of Osmundaceae and its fossil relatives is very complex for such a small group (Bomfleur et al. 2017: fear of "ancestors"?).

[Hymenophyllales, Gleicheniales [Schizaeales [Salviniales [Cyatheales + Polypodiales]]]]: protostele +; sporangia in sori, annulus ± oblique, continuous; loss of chloroplast trnK-UUU, trnT-UGU, trnS-CGA.

Age. (297-)286(-272) Ma is the age for this node in Y. L. Qiu et al. (2007), ca 280.1 Ma in Schuettpelz and Pryer (2009: Hymenophyllales sister to the rest), (306.8-)278.7(-252.3) Ma in Rothfels et al. (2015b) and ca 273 Ma in Hennequin et al. (2008).

Evkaikina et al. (2016) suggest that all other monilophytes have stems with a single apical cell, Marattiales and Osmundales have several such cells.

[Hymenophyllales + Gleicheniales] [if a clade]: ?

Age. The age for this node in Y. L. Qiu et al. (2007) was estimated at (283-)273(-259) Ma and in Rothfels et al. (2015b) at (276.7-)237.2(-192.4) Ma.

HYMENOPHYLLALES A. B. Frank

Just the one family, 9 genera, 435 species.

HYMENOPHYLLACEAE Martinov

Epiphytes common; mycorrhizae uncommon; (roots 0); axillary buds +; petiole vasculature ?; fronds 1 cell thick between veins, stomata 0; sporangia on ± elongated receptacle, maturation basipetal; spores globose, green; gametophyte filamentous or ribbon-like, mycorrhizae 0; embryo not with tetrad/octant formation; x = 36; rpl23-trnI-trnL gene sequence in the LSC.

9 [list]/435. Especially tropical, south temperate

Age. Crown Hymenophyllaceae are estimated to be (191.7-)173.6(-155.1) Ma (Hennequin et al. 2008), (194.7-)176.2(-156.5) Ma (Del Rio et al. 2017), (190.4-)185.1(-174.7) Ma (Schuettpelz & Pryer 2009) or ca 243 Ma (Testo & Sundue 2016).

The earliest fossils of the family (Hopetedia praetermissa, stem Hymenophyllaceae) are from the Late Carnian (Jurassic, ca 230 Ma) of North Carolina (Axsmith et al. 2001), while fossils assignable to Hymenophyllum are known from the Early Cretaceous of Mongolia (Herrera et al. 2017).

Subfamilies: Hym. (107.9-)85(-62.1) Ma, Tri. (159.6-)144(-128.4) Ma (Hennequin et al. 2008) although there is a lot of scatter - uncertainties in topology and different methods of analysis sometimes give very different results.

Evolution: Divergence & Distribution. For ages of clades in the family, see Hennequin et al. (2008). Diversification in Trichomanes is estimated to have begun in the middle of the Jurassic and that in Hymenophyllum in the middle of the Cretaceous (Schuettpelz & Pryer 2007), ca 147.3 versus ca 41.9 Ma (Schuettpelz & Pryer 2009), or ca 90. 5 versus 49.8 Ma (Del Rio et al. (2017); see also Schuettpelz and Pryer (2007) for other dates.

Del Rio et al. (2017) discuss Hymenophyllum on New Caledonia, where the ten (out of 17) species that they sampled represented eight dispersal events; they probably came from New Zealand.

For gametophyte variation and evolution in Hymenophyllaceae, see Dassler and Farrar (1997).

Schuettpelz and Pryer (2007) discuss the rate of molecular evolution of Hymenophyllaceae, with an apparent slow-down in Hymenophyllum itself. Kuo et al. (2018b) outline the extensive variation in the chloroplast genome, with i.a. rearrangements in the long single copy region (see also Ruíz-Ruano et al. et al. 2018).

Ecology & Physiology. Around half the family is epiphytic (Zotz 2013), and there are also climbing taxa (see Dubuisson et al. 2009 for growth forms in Hymenophyllum). Epiphytism in Trichomanes evolved before that in Hymenophyllum, the plants probably growing on the stems of Cyatheaceae on which species of Trichomanes are still often to be found (Hennequin et al. 2008; see also Lehnert & Krug 2019). The epiphytic habit is likely to have been ancestral in Hymenophylloideae (Hennequin et al. 2008; Schuettpelz & Pryer 2009), while in Trichomanoideae the terrestrial habit is more common, but they include a substantial clade of epiphytic taxa. When the epiphytic habit evolved is rather unclear, but some time between the early Mid Jurassic and the later Cretaceous seems likely (Hennequin et al. 2008). Lehnert and Krug (2019) emphasized that in the family as a whole the epiphytic taxa are low epiphytes, that is, they grow on the trunk and lower branches and quite close to the ground (the story in Lehnert et al. 2017 is rather complex). Despite the apparently delicate nature of fronds of Hymenophyllaceae, dessication tolerance is at least sometimes well developed - rather like mosses (Proctor 2003, 2012). Indeed, the sporophytes of some epiphytic trichomanoid ferns have lost both cuticle and roots ("regressive evolution" - Dubuisson et al. 2011), and may be functionally similar to bryophytes; the stem stele may have just a single vascular element (Dubuisson et al. 2013; see also Dubuisson et al. 2003b).

Bacterial/Fungal Associations. The filamentous or ribbon-like gametophytes and rootless sporophytes - and even their rhizomes - are thin and wiry, and this, associated with the epiphytic habitat, means that mycorrhizae are at most uncommon here (Pressel et al. 2016). Lehnert and Krug (2019) noted both that dark septate endophytes were common in Hymenophylloideae, possibly the ancestral condition for the clade, while in Trichomanoideae arbuscular mycorrhizal associations are also much more common (Lehnert & Krug 2019: optimization seems rather erratic).

Phylogeny. For the phylogeny of the family, see Pryer et al. (2001b) and Dubuisson et al. (2003a, 2013), for that of Trichomanes and relatives, see Ebihara et al. (2007), for that of Hymenophyllum, with a focus on the large subgenus Mecodium, which turns out to be polyphyletic but common in a number of basal clades, see Hennequin et al. (2006), and for a possible base chromosome number in the family - previous suggestions x = 6-9, 11, 13, but here = 36 - see Hennequin et al. (2010).

[Gleicheniales + The Rest] (if the clade exists): (spores monolete, perine closely attached to exine).

Age. The stem group age of this node is ca 276.4 Ma (Schuettpelz & Pryer 2009).

GLEICHENIALES Schimper

Root steles with 3-5 protoxylem poles; rhizome with scales; frond veins anastomosing; sporangium maturation simultaneous; (gametophyte axial - Stromatopters), (rhizoids septate); antheridia with 6-12 narrow curved or twisted cells in walls; x = 20, ... 116, nuclear genome [1C] 2.96 pg [?sampling]. 3 families, 10 genera, 172 species.

Age. Crown-group Gleicheniales are ca 262.2 Ma (Schuettpelz & Pryer 2009) or (252.4-)196.1(-134) Ma (Rothfels et al. 2015b).

Ecology & Physiology. Aluminium accumulation is quite common in Gleicheniales (Schmitt et al. 2017).

Synonymy: Dipteridales Doweld, Matoniales Reveal, Stromatoperidales Reveal

Includes Dipteridaceae, Gleicheniaceae, Matoniaceae

GLEICHENIACEAE C. Presl

Leaves indeterminate, pseudodichotomously forked (not - Stromatopteris); petiole vasculature incurved U-shaped; spores (bilateral), monoulcerate; (embryo exoscopic, cell wall vertical, gametophyte (axial, subterranean, mycorrhizal), with clavate hairs; x = 22, 34, etc.. 6 [list]/160.

There has been a chloroplast genome inversion in the family (Wolf & Roper 2008).

Stromatopteris has a mycoheterotrophic gametophyte, but the identity of the fungus is not known (Merckx et al. 2013a; Imhof et al. 2013).

[Dipteridaceae + Matoniaceae]: ?

DIPTERIDACEAE Seward & E. Dale

Petiole vasculature inverted Ω-shaped bundle; frond veins reticulate, areoles with free-included veins, density 4.4-5.6 mm/mm2; sporangia with "long" stalks, (spores bilateral, monolete); x = 33. 2 [list]/11. N.E. India to N.E. Australia, earlier in Caenozoic widespread.

For relationships in the family, including its fossil relatives, see Choa and Escapa (2017). Inc. Cheiropleuriaceae.

MATONIACEAE C. Presl

Stems solenostelic, with two vascular cylinders and a central bundle; petiole vasculature ± inverted Ω-shaped bundle; fronds or pinnae ± dichotomously branched; sporangia in ring surrounding central "receptacle", sorus indusiate; x = 25, 26. 2 [list]/4. Malesia, previously widespread.

[Schizaeales [Salviniales [Cyatheales + Polypodiales]]]: plant with hairs; endospore 2-layered; antheridium wall ca 3 cells across; chloroplast genome with two overlapping inversions.

Age. (281-)266(-250) Ma is the age for this node in Y. L. Qiu et al. (2007), ca 264.6 Ma in Schuettpelz and Pryer (2009) and (289.4-)258.3(-235.2) Ma in Rothfels et al. (2015b).

SCHIZAEALES Schimper

Fronds differentiated into fertile/sterile portions [hemidimorphic]; petiole vasculature?; sporangia on leaf segments lacking laminar tissue, annulus sub-apical, transverse, continuous; n = 28-504. - 3 families, 4 genera, 198 species.

Age. The crown-group age of Schizaeales is estimated to be around 218.4 (Schuettpelz & Pryer 2009) or 183.6 Ma (Hennequin et al. 2008).

Includes Anemiaceae, Lygodiaceae, Schizaeaaceae

LYGODIACEAE M. Roemer

Fronds indeterminate, climbing, pseudodichotomously branched, with a bud in angle of branch; one sporangium/sorus, subtended by antrorse indusium-like flange; x = 29, 30.

1 [list]/25-40. Tropical and warm temperate.

[Anemiaceae + Schizaeaceae]: sporangia not in sori, exindusiate.

ANEMIACEAE Link

Rhizome with dictyostele or solenostele, (with pockets axillary to the fronds); spores tetrahedral, with parallel solid ridges (ridges hollow, centre spongy); x = 38.

1 [list]/130. Africa, to the Mascarenes.

For a largely well-resolved phylogeny of the family and optimisation of characters on to the tree, see Labiak et al. (2015: spore morphology!).

SCHIZAEACEAE Kaulfman

Inner pericyclic cells 6, 8, thickened; fronds undivided or fan-shaped, veins dichotomizing; sporangia borne on marginal projections at blade tip; spores monolete, bilateral, smooth; gametophyte filamentous, (white, subterranean, tuberous or filamentous), rhizoids septate, (borne 2-3 together on large, vacuolated cells, = rhizoidophores),(embryo exoscopic, cell wall vertical); chloroplast ndh genes lost, small single copy reduced in size; x = 77, 94, 103.

2 [list]/35. Pantropical to temperate.

For the chloroplast genome of Schizaeaceae, see Labiak and Karol (2017). There has been an inversion in the chloroplast genome somewhere around here (Wolf & Roper 2008); see also the next node up.

Actinostachys has a mycoheterotrophic gametophyte, but it is not known what the fungus is (Merckx et al. 2013a; Imhof et al. 2013).

[Salviniales [Cyatheales + Polypodiales]]: sporangium stalk 1-3 cells across [?position]; nuclear genome duplication [?here]; two [more!] overlapping inversions in chloroplast genome.

Age. The age of this node is estimated to be around 234.7 Ma (Schuettpelz & Pryer 2009) and (269.1-)231.6(-190.8) Ma (Rothfels et al. 2015b).

Genes & Genomes. The rate of evolution of the 18S nuclear gene was lower than in the other vascular plants examined (Stenøien 2008: lycophytes not included). For a possible nuclear genome duplication in this area, see F.-W. Li et al. (2018) and F.-W. Li et al. (2018).

SALVINIALES Link

Aquatics, mycorrhizal fungi 0; roots 0; aerenchyma +; veins ± anastomosing; sterile/fertile frond dimorphism; plant heterosporous, sporangia lacking annulus; megaspore single; gametophyte development endosporic; nuclear genome [1C] ca 2.38 pg; nrDNA with 5.8S and 5S rDNA in separate clusters. - 2 families, 5 genera, 82 species.

Age. Crown-group Salviniales are estimated to be around 158.9 Ma (Hennequin et al. 2008), ca 186.8 Ma (Schuettpelz & Pryer 2009) or (204.5-)153.5(-150) Ma (Testo & Sundue 2016).

See Nagalingum et al. (2006, 2007: sporocarp structure) and Nagalingum et al. (2008: phylogeny).

Synonymy: Marsileales von Martius, Pilulariales Berchtold & Presl

Includes Marsileaceae, Salviniaceae

MARSILEACEAE Mirbel

Fronds simple, linear, or to 4 divisions/frond; sori in stalked bean-shaped sporocarps [folded pinnae]; megaspore with acrolamella over the exine aperture, perine gelatinous; female gametophyte with 1 archegonium; n = 10, 19, 20, nuclear genome [1C] 0.83-1.34 Gb.

3 [list]/61. Tropical to temperate.

For the phylogeny of Marsilea and character evolution there, see Nagalingum et al. (2007).

Synonymy: Pilulariaceae de Candolle

SALVINIACEAE Martinov

Plant free-floating; fronds sessile, 2-ranked, <2.5 cm long, simple; sporophyte-gametophyte boundary smooth [?level]; n = 9, 22, nuclear genome [1C] 0.25[Salvinia]-0.75 Gb.

2 [list]/21. Tropical to warm temperate.

Age. The crown-group age of this clade may be around 89 Ma (Hennequin et al. 2008).

Synonymy: Azollaceae Wettstein

Evolution: Ecology & Physiology. Azolla has the cyanobacterium Nostoc in its tissues and is an important nitrogen fixer in rice paddies, etc. (Warshan et al. 2018 for the relationships of N-fixing members of Nostoc - N. azollae is unrelated to the other fixers). Fronds of A. filiculoides contain N. azollae (the cyanobacterium is vertically transmitted), and massive amounts of biomass are produced - ca 39 t ha-1 yr-1, around one quarter of which is protein; gram negative rhizobial bacteria are also very common in the leaf pockets where Nostoc lives, but they do not fix nitrogen (Dijkhuizen et al. 2018). The symbiosis is some 100 Ma, there apparently has been cospeciation, but with one host switch, and there has been loss or pseudogenization of some housekeeping genes in Nostoc (F.-W. Li et al. 2018). At times in the Eocene ca 50 Ma Azolla covered vast areas of the Arctic Ocean during periods of lower salinity ((Brinkhuis et al. 2006).

Salvinia produces two floating leaves, a bud, and a submerged leaf at each node, and as the bud develops, it produces additional buds... (Leomoin &Posluszny 1997).

[Cyatheales + Polypodiales]: dictyostele +; hydathodes +; IR with several large inversions, ycf2 duplication.

Age. The age for this node is ca 211 Ma in Y. L. Qiu et al. (2007), ca 223.2 Ma in Schuettpelz and Pryer (2009), (270.1-)228.8(-187.5) Ma in B. Zhong et al. (2014b) and (238.1-)204.6(-179) Ma in Rothfels et al. (2015b).

Evolution: Ecology & Physiology. Aluminium accumulation is quite common here (Schmitt et al. 2017).

CYATHEALES A. B. Frank

Petiole vasculature complex, ± inverted Ω-shaped; hairs +; sori terminal on veins, indusiate, indusium with outer and inner parts; sporangium stalk ca 5 cells across, annulus oblique; antheridium walls ³5 cells across; nuclear genome [1C] ca 7.91 pg [?sampling]. - 8 families, 13 genera, 745 species.

Age. The crown-group age of Cyatheales is ca 186.7 Ma (Schuettpelz & Pryer 2009) or rather younger, (167.8-)109.1(-56.8) Ma (Rothfels et al. 2015b: Alsophila sister to the rest).

There seems to be a slow-down in the rate of evolution at the base of Cyatheales (Rothfels et al. 2015b and references).

Synonymy: Dicksoniales Reveal, Hymenophyllopsidales Reveal, Loxsomatales Reveal, Metaxyales Doweld, Plagiogyriales Reveal

Includes Cibotiaceae, Culcitaceae, Cyatheaceae, Dicksoniaceae, Loxsomataceae, Metaxyaceae, Plagiogyriaceae, Thyrsopteridaceae

THYRSOPTERIDACEAE C. Presl

Indusium cup-shaped, receptacle columnar, clavate; n = ca 78.

1 [list] /1: Thyrsopteris elegans. Juan Fernandez, fossils widespread.

[[Loxsomataceae [Culcitaceae + Plagiogyriaceae]] [Cibotiaceae + Cyatheaceae + Dicksoniaceae + Metaxyaceae]]: ?

Age. The age of this clade is around 183.1 Ma (Hennequin et al. 2008).

[Loxsomataceae [Culcitaceae + Plagiogyriaceae]]: ?

LOXSOMATACEAE C. Presl

Petiole vasculature complex; indusium urceolate, receptacle elongate, often exserted; gametophyte with scale-like hairs; n = 46, 50.

2 [list]/2. Scattered, Costa Rica to Bolivia, New Zealand.

[Culcitaceae + Plagiogyriaceae]: ?

CULCITACEAE Pichi Sermolli

Petiole vasculature inverted Ω-shaped; outer indusium scarcely differentiated; sori with paraphyses; n = 66.

1 [list]/2. ± Tropical, scattered, to New Caledonia, also Azores to the Iberian Peninsula, not Africa.

PLAGIOGYRIACEAE Bower

1 root/petiole; petiole with pneumatophores (0), vasculature U- or V-shaped bundle, (three elliptic meristeles); fronds fimorphic; young fronds with dense, pluricellular, mucilage-secreting hairs; indusium 0, frond margin recurved [= "false indusium"]; n = 62, 64-66, 75.

1 [list]/15. Eastern India and Japan to the Solomon Islands, Mexico to Brazil, the Antilles, often montane.

Chemistry, Morphology, etc.. See X.-C. Zhang and Noteboom (1998: revision) and R-x. Wang et al. (2018: cytology).

[Cibotiaceae + Cyatheaceae + Dicksoniaceae + Metaxyaceae]: petiole vasculature with multiple traces coming from a ± U-shaped bundle; paraphyses +.

Age. This clade may be some 159 Ma (Hennequin et al. 2008).

CIBOTIACEAE Korall

Stomata with three subsidiary cells; spores with equatorial flange, usu. parallel ridges on distal face; n = 68.

1 [list]/9. Tropical, Central America, Southeast Asia to Malesia, Hawaii.

[Cyatheaceae + Dicksoniaceae]: the crown-group age of this clade, if it exists, is ca 150 Ma (Janssen et al. 2008) or ca 157 Ma (Noben et al. 2017).

CYATHEACEAE Kaulfman

Stem with polycyclic dictyostele; fronds with multiple traces coming from a U-shaped bundle; scales large (also small); fronds large; indusium 0 to completely surrounding sporangia); n = 69.

3 [list]/645: Alsophila (275), Cyathea (265), Sphaeropteris (105). Pantropical (to subAntarctic - Alsophila).

Age. Crown-group Cyatheaceae are estimated to be ca 96 Ma (Janssen et al. 2008).

The rate of chloroplast genome evolution has slowed down considerably here, probably because of the long generation time of tree ferns (P. Soltis et al. 2002; esp. B. Zhong et al. 2014b).

Korall and Pryer (2014) outline major biogeographic patterns in the group; initially Gondwanan vicariance seems to be involved, although crown-goup diversification did not begin until ca 96 Ma with subsequent rather limited (for ferns) transoceanic long distance dispersal. The 100+ endemic species of Cyathea on Madagascar may represent a Pliocene diversification (4.24-)3.07> Ma of three separate clades each of which has a fairly lengthy sojourn on the island - thus one Malagasy clade hung around for 30 My before diversification (Janssen et al. 2008, see also Korall & Pryer 2014; Test & Sundue 2016 for very high speciation and extinction rates). Bystriakova et al. (2011) discussed niche evolution.

See Korall et al. (2006, 2007) for a phylogeny.

[Dicksoniaceae + Metaxyaceae]: the crown-group age of this clade, if it exists, is (172-)132(-94.5) Ma (Testo & Sundue 2016).

DICKSONIACEAE M. R. Schomburgk

Plants not Al accumulators; adaxial [outer!?] valve of sorus formed by reflexed frond segment margin and often differently coloured from the other; n = 56, 65.

3 [list]/35. Tropical America, St Helena, Malesia to the Antipodes and New Caledonia.

Age. The crown age of Dicksoniaceae is ca 135 Ma (Noben et al. 2017).

Phylogeny. For the phylogeny of Dicksoniaceae, with Calochlaena sister to the rest, see Noben et al. (2017). There is quite a rich fossil record of the family (but not of Calochlaena) from ex-Gondwanan continents and the family has a basically Gondwanan distribution today, but the three genera have diversification fuses of 80-110 Ma, and some species of Dicksonia have estimated ages older than the islands they currently inhabit, so understanding the biogeography of the family is tricky (Noben et al. 2017).

Synonymy: Lophosoriaceae Pichi-Sermolli

METAXYACEAE Pichi Sermolli

Indusium 0; n = 95, 96.

1 [list]/6. Tropical South America.

POLYPODIALES Link

Roots black, wiry [?level]; rhizome dorsiventral [?level]; petiole vasculature with a single bundle; sporangial maturation mixed; stalk 1-3 cells thick, annulus vertical, interrupted by stalk and stomium; neochrome/phy 3 +. -

Age. This clade is estimated to be around 260 Ma (Testo & Sundue 2016), (200-)176(-163) Ma (H. Schneider et al. 2004a), ca 191 Ma (Schuettpelz & Pryer 2009), (220.1-)184.2(-149.2) My (Rothfels et al. 2015b) or around 150 Ma (Hennequin et al. 2008).

Synonymy: Aspleniales Reveal, Athyriales Schmakov, Blechnales Reveal, Dennstaedtiales Doweld, Dryopteridales Schmakov, Lindseales Doweld, Negripteridales Reveal, Parkeriales A. B. Frank, Platyzomatales Reveal, Pteridales Doweld, Thelypteridales Doweld

For a discussion of characters for this clade, particularly the eupolypod II group, see Sundue and Rothfels (2014).

Phylogeny. Pteridaceae and Dennstaediaceae were well supported as successive sister taxa to the eupolypods (Lehtonen 2011; Rothfels et al. 2015b; J.-M. Lu et al. 2015; Qi et al. 2018). Within Polypodiales, Lindsaeaceae are probably sister to all others (see also Rothfels et al. 2015b), but the genera Cystodium, ex Dicksoniaceae, and Lonchitis and Saccoloma, both ex Dennstaediaceae and all as separate families below - are also in this area (Lehtonen 2011; Lehtonen et al. 2012; Qi et al. 2018). Structural changes in the chloroplast genome (Wolf & Roper 2008; Wolf et al. 2010, 2011) support relationships that are largely consistent with those suggested by sequence analyses, while H. Shen et al. (2017) suggest relationships based on analyses of transcriptome data.

Within the eupolypod I clade, Lehtonen (2011) found that Polypodiaceae were not monophyletic in parsimony analyses, and other relationships were unclear. For more on relationships in the eupolypod I clade, see L.-B. Zhang and Zhang (2015); a couple more families may still be needed. Thus H.-M. Liu et al. (2014), as they placed the odd genera Pteridrys and Pleocnemia in Tectariaceae and Dryopteridaceae respectively, recognised an [Arthropteridaceae + Tectariaceae] clade (the former is synonymized under the latter here). Further work on the Tectariaceae area (see Dong et al. 2018; X.-M. Zhou et al 2018) has indeed suggested both more genera and new families in this area.

Within the eupolypod II clade, Lehtonen (2011) found that relationships were mostly poorly supported, even if the families themslves were largely monophyletic. Cystopteris and relatives form a clade that may be sister to the rest (Rothfels et al. 2009, esp. 2012a, 2013, 2015b: Qi et al. 2018). However, relationships between the members of the other eupolypod II groups sampled by Qi et al. (2018) are rather different from those below - the relationships [[Diplaziopsidaceae + Aspleniaceae] [Rhachidosoraceae [Thelypteridaceae [Woodsiaceae [Athyriaceae [Blechnaceae + Onocleaceae]]]]]] were often recovered, although the basal grouping [[Diplaziopsidaceae + Aspleniaceae] [Cystopteridaceae + the rest]] was also found.

Classification. Currently relationships are a bit uncertain around here and a number of small families are being described, but the links below lead to the families as recognized by The Pteridophyte Phylogeny Group (2016) - not that there is currently much below for most of them...

Includes Aspleniaceae, Athyriaceae, Blechnaceae, Cystodiaceae, Cystopteridaceae, Davalliaceae, Dennstaedtiaceae, Desmophlebiaceae, Didymochlaenaceae, Diplaziopsidaceae, Dryopteridaceae, Hemidictyaceae, Hypodemataceae, Lindsaeaceae, Lomariopsidaceae, Lonchitidaceae, Nephrolepidaceae, Oleandraceae, Polypodiaceae, Pteridaceae, Rhacidosoraceae, Saccolomataceae, Tectariaceae, Thelypteridaceae, Woodsiaceae

[Lindsaeaceae, Lonchitidaceae, Saccolomataceae, Cystodiaceae]: ?

LINDSAEACEAE M. R. Schomburgk

Innermost cortical layer of root usu. of 6 large cells; stele protostelic, with internal phloem; petiolar vasculature V-shaped and then with two bundles; indusium opening towards margin; x = 34, 38, etc.

7 [list]/235: Lindsaea (180). Pantropical (subtropical).

See Lehtonen et al. (2010) for a phylogeny and generic classification.

Age. (193.6-)165.4(-113.7) Ma is a suggestion for the age of this node made by Rothfels et al. (2015b).

CYSTODIACEAE J. R. Croft

Petiole vasculature with two complex bundles.

1 [list]/1: Cystodium sorbifolium. Malesia.

LONCHITIDACEAE Doweld

Petiole vasculature with two bundles.

1 [list]/2. Tropical America and Africa, Madagascar.

SACCOLOMATACEAE Doweld

Scales?; petiole vasculature with inverted Ω-shaped bundle; spores also with distinctive ± parallel branched ridges; x = ca 63.

1 [list]/18. Tropical, not mainland Africa.

[Dennstaedtiaceae + Pteridaceae]: SiO2 accumulation common; petiole vasculature with inverted Ω-shaped bundle.

Age. B. Zhong et al. (2014b) suggested an age of (217-)154.3(-93.1) Ma for this node, Schuettpelz and Pryer (2009) an age of ca 110.8 Ma, and H. Schneider at al. (2016) ages somewhere around 137.4-95.6 Ma.

DENNSTAEDTIACEAE Lotsy

Chimaeric red/far red light photoreceptor [phy 3, neochrome]; stele?; hairs jointed; petiole bearing buds; x = 26, 29.

10 [list]/265: Hypolepis (80), Dennstaedtia (70), Microlepia (60).

Age. Crown-group Dennstaedtiaceae are estimated to be ca 72.2 Ma (Schuettpelz & Pryer 2009) or around 98-66.2 Ma (H. Schneider et al. 2016).

PTERIDACEAE E. D. M. Kirchn.

(Epiphytic), (xeric); indusium 0; (spores bilateral); (gametophyte ribbon-like, mycorrhizal fungi 0); x = 29, 30.

53 [list]/>1,210: Pteris (250), Adiantum (225), Cheilanthes (100), Jamesonia (50), Myriopteris (45), Aleuritopteris (40), Antrophyum (40), Haplopteris (40), Pellaea (40). Worldwide.

Age. This clade is around 106.3 Ma (Schuettpelz & Pryer 2009), 90 Ma (Rothfels & Schuettpelz 2014), or about 99.9-87.4 Ma (H. Schneider et al. 2016).

Kramopteris resinatus, in amber ca 100 Ma from Myanmar, is placed at the split between Monachosorum and other Pteridaceae (H. Schneider et al. 2016).

Evolution: Divergence & Distribution. For Adiantum in the West Indies, the result of at least 17 colonization events and some subsequent speciation in the older islands, see Regalado et al. (2018).

Ecology & Physiology. Cheilanthoid ferns, some 400 or more species, can grow in very dry conditions (e.g. Grusz et al. 2014 and references). Kao et al. (2019) discuss the evolution of farina, whitish and powdery, made up of a variety of lipophilic flavonoid aglycones produced by glands on the margins of the gametophytes and the lower surface of the fronds - the focus is on Notholaena, a denizen of deserts of southwest North America. The pattern of gain (?and loss) of this feature is complex, in part independent in the two generations, and there are some species that lack farina but seem to do perfectly well growing in the same conditions as their farina-producing relatives (Kao et al. 2019).

Vittarioid ferns are commonly epiphytic and are one of the major monilophyte epiphytic groups. Twig epiphytes commonly lack mycorrhizae, but mycorrhizae are more common in low epiphytes where the plants grow quite close to the ground (Lehnert et al. 2016).

Phylogeny. For phylogenies, see Crane et al. (1995), Prado et al. (2007), Schuettpelz (2007), L. Zhang et al. (2016b: Pteridoideae, 14 genera) and Schuettpelz et al. (2016: vittarioids). Pteris has been the subject of some attention, e.g. Chao et al. (2014: position of P. longifolia, the type, unclear) and L. Zhang et al. (2015: Pteris somewhat expanded), while L. Zhang and Zhang (2017) produced a comprehensive phylogeny and infrageneric classification, most infrageneric taxa being well supported. Pryer et al. (2016) found that Adiantum is monophyletic and is sister to Vittaria and its relatives. For some increased rates of molecular evolution, see Rothfels and Schuettpelz (2014). Generic limits are difficult in cheilanthoid ferns (Grusz et al. 2014; Yesilyurt et al. 2015).

Inc. Parkeriaceae Hooker

Eupolypods: petiole vasculature V-shaped bundle, then two bundles, circum-endodermal band + (0); fronds to 1.5 times pinnate; spores monolete, reniform, perine distinct; x = 41.

Age. This clade, which includes most ferns, has been estimated to be (124-)105(-91) Ma (H. Schneider et al. 2004a) or ca 116.7 Ma (Schuettpelz & Pryer 2009).

The oldest eupolypod fossil is from Burmese amber 98 myo - or somewhat more; it was compared with Thelypteridaceae (Regalado et al. 2017), the crown-goup age of which is some 68.5 Ma (Schuettpelz & Pryer 2009).

Phylogeny. Qi et al. (2018) included members of all eupolypod I families in their analyses, and the following relationships were frequently recovered - [Hypodematiaceae [Didymochlaenaceae [Dryopteridaceae [Lomariopsidaceae [Nephrolepidaceae [Tectariaceae [Oleandraceae [Davalliaceae + Polypodiaceae]]]]]]]].

[Didymochlaenaceae [Hypodematiaceae [[Nephrolepidaceae + Lomariopsidaceae] [Dryopteridaceae [Tectariaceae [Oleandraceae [Davalliaceae + Polypodiaceae]]]]]]] / eupolypod I: (plants epiphytic); rhizome scales persistent, dense; petiole vasculature complex, circular in t.s., also with abaxial semicircular arc of smaller bundles; perispore with thick tuberculate folds/wings.

Age. The age of this node is ca 98.9 Ma (Schuettpelz & Pryer 2009) or (172.5-)170(-158.5) Ma (Testo & Sundue 2016).

This is a largely epiphytic clade (e.g. Schuettpelz & Pryer 2009; Tsutsumi & Kato 2006). For rhizome scales, perhaps protecting the plant against dessication and aiding in the absorbtion of water and nutrients, see Tsutsumi and Kato (2008); if they are not at this position on the tree, they should be placed at the next node up (with parallel evolution within Dryopteridaceae).

Phylogeny. For discussion of relationships in the eupolypod I clade, see above.

Within the eupolypod II clade, Cystopteris and relatives form a clade that may be sister to the rest (Rothfels et al. 2009, esp. 2012a, 2013, 2015b: Qi et al. 2018). However, relationships between the members of the other eupolypod II groups sampled by Qi et al. (2018) are rather different from those below - the relationships [[Diplaziopsidaceae + Aspleniaceae] [Rhachidosoraceae [Thelypteridaceae [Woodsiaceae [Athyriaceae [Blechnaceae + Onocleaceae]]]]]] were often recovered, although the basal grouping [[Diplaziopsidaceae + Aspleniaceae] [Cystopteridaceae + the rest]] was also found.

DIDYMOCHLAENACEAE L.-B. Zhang & L. Zhang

Rhizome erect, subarborescent; sori hippocrepiform, somewhat elongated. 1 [list]/?1. ± Pantropical.

[Hypodematiaceae [[Nephrolepidaceae + Lomariopsidaceae] [Dryopteridaceae [Tectariaceae [Oleandraceae [Davalliaceae + Polypodiaceae]]]]]]: ?

HYPODEMATIACEAE Ching

(n = 40).

2 [list]/22.

[Nephrolepidaceae + Lomariopsidaceae]: ?

NEPHROLEPIDACEAE Pichi Sermolli

?

1 [list]/19. Tropical to subtropical.

If the sister-group relationship with Lomariopsidaceae holds, the two are best merged.

LOMARIOPSIDACEAE Alston

; n = 40, 41.

5 [list]/70: Lomariopsis (60).

Phylogeny. For relationships in Lomariopsidaceae and the vicissitudes of the family, see C.-W. Chen et al. (2017).

[Dryopteridaceae [Tectariaceae [Oleandraceae [Davalliaceae + Polypodiaceae]]]]: ?

DRYOPTERIDACEAE Herter

(Epiphytic); fronds >3.5 times pinnate; perine winged; gametophyte strap-like; x = 41.

26 [list]/2,115: Elaphoglossum (600), Polystichum (500), Dryopteris (400), Ctenitis (125), Megalastrum (90), Bolbitis (80), Arachnoides (60), Stigmatopteris (40). World-wide.

Age. Crown-group Dryopteridaceae are around 81.8 Ma (Schuettpelz & Pryer 2009) or (94-)76(-58) Ma (Le Péchon et al. 2016).

Evolution: Divergence & Distribution. Elaphoglossum is the major epiphytic genus in the family - ca 400 species (ca 2/3 of the genus) are epiphytes (Zotz 2013). Diversification within the speciose Polystichum began in the Eocene, (60-)46(-32) Ma (Le Péchon et al. 2016), that in Elaphoglossum began a bit more recently - and that covers about 2/3 of the family. espite the pantropical distribution of Ctenitis, long distance dispersal seems to have been uncommon (Hennequin et al. 2017)

Phylogeny. For a phylogeny of the whole family, see H.-M. Liu et al. (2007, esp. 2015); three subfamilies are recognised, although two genera are unplaced. For a phylogeny of Elaphoglossum, see also Lóriga et al. (2014: 20-15 Ma fossil of ?crown-group Elaphoglossum from Dominican amber). Rouhan et al. et al. (2004) and Vasco et al. (2015) discuss relationships within Elaphoglossum, where the Coast Rican E. amygdalifolia and the Cuban E. wrightii are successively sister to the rest of the genus. Moran et al. (2010a, b) investigate relationships within the bolbitidoid ferns focussing on variation in perispore morphology. Li and Lu (2006a, b), L.-B. Zhang et al. (2012), Sessa et al. (2012a), and McHenry and Barington (2014: exindusiate Andean species monophyletic, sister to Mexican spp.) have been working on relationships within Dryopteris itself, a genus whose limits are being clarified (e.g. Zhang & Zhang 2012) and which shows extensive hybridization at all levels (Sessa et al. 2012b). In Ctenitis, on the other hand, polyploidy is probably rare (Hennequin et al. 2017)Le Péchon et al. (2016) examined phylogenetic and biogeographic relationships in Polystichum, while Labiak et al. (2014) looked at relationships around Lastreopsis, with movement to and fro between Australia and South America towards the middle of the Caenozoic.

Synonymy: Nephrolepidaceae J. Agardh

[Tectariaceae [Oleandraceae [Davalliaceae + Polypodiaceae]]]: frond veins free, parallel or pinnate.

TECTARIACEAE Panigrahi

(Climbers); (rhizome slender), (stipe and pinnae articulated); (frond veins free, parallel or pinnate), with jointed usually stubby hairs; n = 40-42.

7 [list]/270: Tectaria (210). Pantropical, inc. oceanic islands.

Age. Crown-group Tectariaceae (excluding Arthropteris) are around 50.4 Ma (Schuettpelz & Pryer 2009).

Phylogeny. Although in the analysis of H.-M. Liu et al. (2013) the position of Arthropteridaceae was uncertain, F. G. Wang et al. (2014a) found that they were embedded in Polypodiaceae-Tectarioideae, as were Lomariopsidaceae; the latter was in a separate clade in Schuettpelz and Pryer (2009). L.-B. Zhang and Zhang (2014) placed Arthropteridaceae sister to Tectariaceae, but with less than overwhelming support, Lomariopsidaceae were again separate. The circumscription of Tectariaceae as adopted here was also found by L. Zhang et al. (2016a) and the monophyly of Tectaria s. str. confirmed - although three species were somewhat errant in the study of Dong et al. (2018) and were placed in a separate genus, and overall relationships were [Arthropteris [Pteridrys group + Tectariaceae sensu stricto]] - the position of Arthropteris was not that well supported. Indeed, X.-M. Zhou et al. (2018) [Pteridrys group [Arthropteris + Tectariaceae sensu stricto]] - again, support could have been stronger - and recognised all three as families.

Classification. For an infrageneric classification of Tectaria, see L. Zhang and Zhang (2018).

Synonymy: Arthopteridaceae H. M. Liu, Hovenkamp & H. Schneider, Pteridryaceae Li Bing Zhang, X. M. Zhou, Liang Zhang & T. N. Lou

[Oleandraceae [Davalliaceae + Polypodiaceae]]: fronds abscising from rhizome.

OLEANDRACEAE Pichi Sermolli

Fronds abscising just above the base [so leaving phyllopodia], often simple; x = 41.

1 [list]/15. Tropical.

[Davalliaceae + Polypodiaceae]: epiphytes predominant; perispore thin, verrucate, granulate, etc..

Age. The two diverged ca 42 Ma (Sundue et al. 2015).

DAVALLIACEAE M. R. Schomburgk

Fronds >3.5 times pinnate; spores warty, warts close, not constricted at their bases [= verrucate-colliculate]; x = 40.

1 [list]/65. South East Asia to Oceania, western Mediterranean and tropical Africa, both one species.

Age. Crown-group Davalliaceae are ca 19.4 Ma (Schuettpelz & Pryer 2009).

F.-G. Wang et al. (2014b) optimize spore morphology on a phylogeny of the family. Tsutsumi et al. (2016) found seven major clades in the family (one monotypic), but relationships rather depended on the marker analysed. For a generic classification, see Kato and Tsutsumi (2008), however, Tsutsumi et al. (2016: sectional classification) suggest that only a single genus should be recognised, in part because of these phylogenetic uncertainties.

POLYPODIACEAE J. Presl & C. Presl

(Rhizome polysmmetrical); (petiole with one or two vascular bundles - grammitids); indusium 0; (spores green, globose-tetrahedral, trilete - grammitids); (gametophyte strap-like); x = 35-37.

60 [list]/1,650: Oreogrammitis (155), Selliguea (105), Pleopeltis (90), Prosaptia (87), Lepisorus (80), Calymmodon (65), Pyrrosia (65), Leptochilus (50-100), Aglaomorpha (50), Campyloneuron (50), Lellingeria (50), Grammitis (40), Microsorum (40), Pecluma (40), Polypodium (40), Serpocaulon (40), Tomophyllum (40).

Age. Crown-group Polypodiaceae (Loxogramme + The Rest) are about 55.8 Ma (Schuettpelz & Pryer 2009).

Evolution: Divergence & Distribution. For ages of grammitid ferns, see Schuettpelz and Pryer (2009), Sundue et al. (2015) and Bauret et al. (2017). Bauret et al. (2017) note much long distance dispersal from Madagascar to Africa, little in the reverse direction; grammitid ferns may have originated in the New World, with several dispersals across the Atlantic. Diversification in Polypodiaceae may be associated with the uplift of the Andes, clades with broad elevational ranges diversifying faster, speciation being related to climatic/environmental factors, less obviously associated with morphological features (Sundue et al. 2015: epiphytes in general; Kreier et al. 2008: Andean Serpocaulon; L. Wang et al. 2012: Qinghai-Tibet Lepisorus).

Weber and Agrawal (2014) suggested that the evolution of extra-floral nectaries in Pleopeltis was associated with an increase in diversification. Within Selliguea there is some major geographic structuring (He et al. 2018). Janssen et al. (2005) discussed the evolution of the diverse frond morphologies in Drynaria s.l..

Ecology & Physiology. Ca 87% of the species of Polypodiaceae, the grammitid ferns, are epiphytic (Zotz 2013), making them the major epiphytic clade in the monilophytes. However, in the sporophytes here an apparently secondary association with mycorrhizal ascomycetes has developed - unlike, say, in epiphytic Hymenophyllaceae, although in both dark septate endophytes tend to be common - and the plants seem to be dependent on this association; they are often rather small and are twig epiphytes (Lehnert et al. 2016).

Lecanopteris is a myrmecophyte commonly found in ant gardens from Malesia to the Pacific (Haufler et al. 2003: phylogeny; Chomicki et al. 2017a). A a number of taxa catch litter, interestingly, mycorrhizae are uncommon in the non-epiphytic taxa which have been characterized as being "non-mycorrhizal nutrient savers" (Lehnert et al. 2016).

Chemistry, Morphology, etc.. For root anatomy, see H. Schneider (1996, 1997), and for petiole anatomy, see Sundue et al. (2014a, esp. b and references).

Phylogeny. For relationships in Polypodiaceae, see Janssen et al. (2007), a phylogeny of microsoroid ferns, see Kreier et al. (2008), for those in grammitids, see Sundue et al. (2010, 2014a, esp. b, 2015 and references: generic changes) and Bauret et al. (2017) - Grammatis, Enterosora, etc., still polyphyletic, for those of sub-Saharan Dryopteris, see Sessa et al. (2017) and for relationships around Pyrrosia, see Zhou and Zhang (2017: probably ancient hybridization; Hovenkamp's sections have held up quite well). Labiak and Moran (2017) looked at relationships in the neotropical and largely epiphytic genus Campyloneurum, and He et al. (2018: chloroplast genes) examined relationships around Selliguea s.l., finding it to be made up of three main clades. L. Zhang et al. (2019: chloroplast genes) looked at relationships in the Old World Leptochilus; basalmost relationships were unclear, and there was quite extensive infraspecific variation/polyphyly in some species, suggesting that the number of species in the genus had been seriously underestimated.

F.-G. Wang et al. (2014a) include a broadened but monophyletic Tectariaceae as a subfamily of Polypodiaceae while He et al. (2018) attempt to bring stability to the Selliguea area by broadening the concept of that genus - clades around there lack morphological apomorphies...

[Cystopteridaceae [[Rhachidosoraceae [Diplaziopsidaceae [Desmophlebiaceae [Hemidictyaceae + Aspleniaceae]]]] [Thelypteridaceae [Woodsiaceae [Athyriaceae [Blechnaceae + Onocleaceae]]]]]] / eupolypod II: petiole vasculature with V-shaped bundle changing to two, ± elongated/crescent-shaped [in t.s.]; rhachis sulcus wall confluent with the costa of pinna; sorus with 7-21 cells/annulus [?level].

Age. This node is ca 103.1 Ma (Schuettpelz & Pryer 2009) or (196-)186(-184) Ma (Testo & Sundue 2016).

For a discussion of possible apomorphies in this clade, see Sundue and Rothfels (2014).

Phylogeny. For discussion of relationships in the eupolypod I clade, see above.

CYSTOPTERIDACEAE Schmakov

Rhizome long-creeping; veins reaching the frond margin; indusium 0 or hood-like; n = 40, etc..

3 [list]/37: Cystopteris (27).

For phylogenetic relationships, see Rothfels et al. (2013). The genera are monophyletic, but there has been very extensive hybridization within Cystopteris and Gymnocarpidium (Rothfels et al. 2014).

[[Rhachidosoraceae [Diplaziopsidaceae [Desmophlebiaceae [Hemidictyaceae + Aspleniaceae]]]] [Thelypteridaceae [Woodsiaceae [Athyriaceae [Blechnaceae + Onocleaceae]]]]]: ?

[Rhachidosoraceae [Diplaziopsidaceae [Desmophlebiaceae [Hemidictyaceae + Aspleniaceae]]]] : sorus on one side of the vein.

RHACHIDOSORACEAE X. C. Zhang

Rhizome scales clathrate; petiole bundels U-shaped; sori on one side of vein; n = 41. 1 [list]/8.

[Diplaziopsidaceae [Desmophlebiaceae [Hemidictyaceae + Aspleniaceae]]]: rhizome scales not clathrate; scales on mature fronds 0.

DIPLAZIOPSIDACEAE X. C. Zhang & Christenhusz

Plant proliferous [producing plantlets by asexual reproduction]; roots pale, fleshy; fronds soft and fleshy; vein endings raised and thickened, forming a submarginal vein; sori elongated, only on one side of vein; n = 40, 41.

2 [list]/4.

[Desmophlebiaceae [Hemidictyaceae + Aspleniaceae]]: ?

DESMOPHLEBIACEAE Mynssen, A. Vasco, Sylvestre, Moran & Rouhan

Rhizomes erect or decumbent; petiole bundles hippocampiform, U-shaped or free; fronds unequal-pinnate, veins free, forming submarginal vein; sori elongated; spores cristate; n = ?

1 [list]/2. Costa Rica to Brazil.

For information on Desmophlebiaceae, see Mynssen et al. (2016).

[Hemidictyaceae + Aspleniaceae]: ?

HEMIDICTYACEAE Christenhusz & H. Schneider

Fronds unequal pinnate, with submarginal collecting vein and margins with broad membranous border; vein endings raised and thickened; n = 31.

1 [list]/1: Hemidictyum marginatum. S. Mexico to S.E. Brazil.

ASPLENIACEAE Newman

Epiphytes common; root pericyclic sclereids with excentric lumina; rhizome scales clathrate; frond trace single, circumendodermal band surrounding trace 0; petiole vasculature with bundles back-to-back, C-shaped, fusing and becoming X-shaped (V- or U-shaped), circumendodermal band 0; frond usu. 2-3 times pinnate, with small clavate hairs, margins decurrent and forming lateral ridge along rhachis; sori on one side of vein; indusia lateral, linear; sporangium stalk 1 cell across in the middle; spores with decidedly winged perine; x = (35) 36 (38) 39.

2 [list]/730: Asplenium (700). Widely distributed.

Age. Crown-group Aspleniaceae are ca 57.7 Ma (Schuettpelz & Pryer 2009).

For the phylogeny of Asplenium, see H. Schneider et al. (2017); some major clades are largely diploid, others are largely polyploid. Asplenium s.l. includes a large number of epiphytic species (Zotz 2013). Helical, non-lignified wall thickenings (c.f. the velamen of monocots) occur in cortical cells of the roots of some Asplenium, mostly epiphytic species (Leroux et al. 2011).

For generic limits, see Bellefroid et al. (2010 and references); Asplenium s. str. is paraphyletic.

[Thelypteridaceae [Woodsiaceae [Athyriaceae [Blechnaceae + Onocleaceae]]]]: (SiO2 accumulation common [?Athy., Onocl.]); frond once-pinnate/pinnatifid.

THELYPTERIDACEAE Pichi-Sermolli

Petiole vasculature with bundles uniting distally into a gutter shape; hairs acicular, whitish or hyaline, on fron, also on surface and/or margins of rhizome scales; first venation of frond lobe/pinnule develops on basiscopic side [catadromous]; n = 27-36.

8-30 [list]/1034: Amauropelta (215), Sphaeostephanos (185), Goniopteris (120), Pneumatopteris (80), Christella (70), Pronephrium (68), Coryphopteris (47).

Age. The age of crown-group Thelypteridaceae is ca 68.5 Ma (Schuettpelz & Pryer 2009).

For a careful evaluation of generic limits, which are best drawn broadly given the extensive generic polyphyly and highly homoplasious "generic" characters, especially within Cyclosorus s.l., see He and Zhang (2012) and also Almeida et al. (2016).

Oliveira et al. (2017) discuss mucilage secretion by uniseriate glandular hairs here- they call the hairs colleters.

[Woodsiaceae [Athyriaceae [Blechnaceae + Onocleaceae]]]: ?

WOODSIACEAE Herter

Plant epipetric; rhizomes suberect; petiole bases persist; circumendodermal band 0; indusium basal, of many strap-shaped or filamentous segments, receptacle raised; n = 33, 38, 39, 41.

1 [list]/39. Mostly montane, northern hemisphere.

[Athyriaceae [Blechnaceae + Onocleaceae]]: ?

Woodsia may be sister to this whole clade (Schuettpelz & Pryer 2009: Fig. S1).

ATHYRIACEAE Alston

Mature frond with abundant anthocyanin; margins entire ["smooth"], petiole base swollen, (starch-containing), ± persistent [= trophopod]; corniculae/scales at adaxial junction of pinna costa with rachis; sori on both sides of vein, indusia opening to face away from vein [two linear back to back sori/J-shaped indusium wrapped around a vein]/(on one side of vein); n = 40, 41.

3 or 6 [list]/650: Diplazium (350), Athyrium (160-230), Deparia (92). Mostly terrestrial, understory.

Age. Crown-group Athyriaceae are around 78.4 Ma (Schuettpelz & Pryer 2009).

For the high diversification rate in the family, highest in the ferns as a whole, see Testo and Sundue (2016), and for divergence dates and biogeography in Diplazium, perhaps a member of the boreotropical flora in the Eocene, see Wei et al. (2015).

Wei et al. (2013) evaluated relationships within Diplazium and found i.a. that species previously assigned to Allantodia in particular were scattered through the tree; they circumscribed Diplazium broadly and provided an infrageneric classification for it. Wei et al. (2017) disentangled relationships in Athyrium, chipping off three small genera, while Kuo et al. (2017, 2018a) looked at relationships in Deparia.

Wei et al. (2013) provided an infrageneric classification for Diplazium and Kuo et al. (2017, 2018a) one for Deparia with the groups morphologically characterized.

[Blechnaceae + Onocleaceae]]: fronds dimorphic [fertile and sterile].

BLECHNACEAE Newman

Young fronds reddish; petiole vasculature complex, abaxially also with three to many round vascular bundles arranged in an arc; (fronds monomorphic), venation catadromous, veins forming narrow areoles near the costa [costular anastomoses]; sori linear, on subcostal commissural vein, (on arches of areolae), (acrostichoid), indusia opening towards costa, (0); perine winged; x = 34, [n = 27-29, 31-37, 40].

24 [list]/265: Parablechnum (65), Austroblechnum (40). Cosmopolitan.

Age. Crown-group Blechnaceae are around 59.8 Ma (Schuettpelz & Pryer 2009), however, an age of almost 100 Ma is suggested by Vicent et al. (2017: also more dates).

For some biogeography, see Moran and Smith (2001) and Vicent et al. (2017). Moran et al. (2018) look at spore morphology in the context of phylogeny.

Perrie et al. (2014) discuss relationships in Blechnaceae, circumscribing Blechnum rather broadly. For a rather splitty classification, with three subfamilies, see de Gasper et al. (2016, 2017: justification p. 440); "traditionally" there were ca 10 genera.

ONOCLEACEAE Pichi-Sermolli

Circumendodermal band 0; petiole basally ± swollen, vascular bundles uniting distally into a gutter shape; trophopods +; fronds dimorphic; sori enclosed by reflexed lamina margins; indusium deltate; sorus with 24-35 cells/annulus; spores chlorophyllous; n = 37, 39, 40.

4 [list]/5. Northern Hemisphere.

Age. Onoclea sensibilis is known fossil from Palaeocene North America 62-58 Ma, the fossils being remarkably similar to extant individuals (Rothwell & Stockey 1991).