Spore Evidence for the Origin of Isoetalean Lycopsids?

A new hypothesis about the origin of isoetalean lycopsids was proposed based on palynological data. The occurrence of three apical papillae on the proximal surfaces of miospores is a significant palynological feature that is clearly defined in both isoetalean and selaginellalean clades. Three apical papillae appeared for the first time within lower Silurian (Wenlockian ca. 430 My) and only in rhyniophytoid plants. Using this observation, we suggest that isoetalean lycopsids could have evolved directly from rhyniophytoids and not from protolepidodendralean lycopsids in the middle Devonian (Eifelian–Givetian) as previously suggested, because protolepidodendralean spores do not possess three apical papillae. Spores with three apical papillae, reported as dispersed as well as in situ, were recorded continuously from the lower Silurian (Wenlockian) through the Devonian, Carboniferous, Permian, Mesozoic to Cenozoic era and form a phylogenetically independent lineage.


Introduction
Lycopsids are a monophyletic group of extant plants with about 1250 species growing globally in all climatic belts [1] and are believed to be the oldest living lineage of vascular plants. It is possible to state that the "golden age" for lycopsids was the Carboniferous era (358.9-298.9 My) especially in coal-forming swamps in tropical wet areas where lycopsids were the dominant plant group and the most important contributors to biomass. The position of sporangium in the axil or on the upper surface of the leaf or sporophyll is one of the most important features. Sporophylls can be among photosynthetic microphylls or (as non-photosynthetic sporophylls) aggregated into cones or fertile zones [2]. The sporangia are on the adaxial side, i.e., upper side of the leaf. Sporophytes are the dominant form. The majority of lycopsids produce one type of spores (homosporous) but some forms yield two types (heterosporous). Some cones possess only one type (monosporangiate) while others have two types of spores (bisporangiate). Zosterophylls were probably a sister group to lycophytes [1]. Today lycopsids are herbaceous but, especially during the Carboniferous, many of them were arborescent in form. Three lycopsid orders, i.e., Lycopodiales, Selaginellales and Isoetales are usually recognized as three independent lineages/clades [1].

Protolepidodendrales
Protolepidodendralean plants were mostly herbs, but sometimes subwoody and small trees that occurred from the Devonian to lower Mississippian. Unlike all other lycophytes however, the Protolepidodendrales bore leaves which were forked at the tips. Protolepidodendralean plants were the first members of the lycophyte lineage to evolve wood and bark, a modified shoot system that acts as a rooting system, bipolar growth and an upright habit [1]. Almost all of them were aligulate [2]. This group underwent a dynamic development during the last 25 years because more than a half of protolepidodendralean taxa have been erected during the last quarter of century, most of them from the late Devonian of China [7]. The most common protolepidodendralean plant genera are Barsostrobus, Hefengiostrobus, Cyclostigma, Hoxtolgaya, Leclercquia and Minarodendron. Protolepidodendralean lycopsids produced micro-and megaspores [7] but none of their spores possessed three apical papillae. Protolepidodendralean miospores are represented mainly by those of the genus Acinosporites and megaspores by those of the genera Lagenicula and Lagenoisporites. Protolepidodendraleans had an important position in phylogenetical scheme [1] because they should be a parent group for the subsequent origin of the selaginellalean and isoetalean clades.

Rhyniopsida
The Rhyniopsida are a very important group of polysporangiate fossil plants that were defined as plants with naked (without emergences) dichotomizing axes with sporangia Life 2023, 13, 1546 3 of 14 that are terminal, mainly fusiform and may dehisce longitudinally; they are diminutive plants and, in so far as is known, have a small terete xylem strand with a central protoxylem [1]. The whole group of rhyniophytoid plants is probably heterogeneous. The oldest unquestionable sporophyte of the earliest vascular land plant is Cooksonia barrandei from the Wenlock (432 Ma) Series in the Czech Republic [8] that yielded trilete crassitate spores of the Aneurospora type. The position of rhyniopsids is crucial for subsequent plant phylogeny ( Table 1). The best-known genus is Cooksonia with several fertile specimens that yielded in situ spores [9]. Generally, it is possible to recognize two main morphological types of rhyniopsid in situ spores: crassitate (genera Ambitisporites, Aneurospora, Streelispora, Synorisporites and Retusotriletes) and non-crassitate (the genus Apiculiretusispora). Rhyniophytoid genera that yielded in situ spores include Cooksonia, Aberlemnia, Concavatheca, Pertonella and Renalia. It is possible to divide rhyniopsida into cooksonioid and renalioid taxa [1]. The genus Rhynia gave the name to the whole group and occurs in the Rhynia cherts, i.e., the Rhynia locality (407 Ma), UK, that represents unique fossilization with animals, fungi, algae and bacteria including the earliest records of plant life cycles. C. pertoni subsp. apiculispora Streelispora newportensis/Aneurospora Lochkovian [11] Caia langii Retusotriletes Přídolí [11] Cooksonia cambrensis Ambitisporites Přídolí [12] Pertonella dactylethra Retusotriletes coronatus Přídolí [12] Renalia hueberi Retusotriletes/Apiculiretusispora Lochkovian [13] Salopella allenii Retusotriletes Přídolí [14]

Material and Methods
The maceration methods used for dispersed and in situ spores mentioned in this paper were described in the papers of cited authors. Typically, hydrochloric acid was used for 5-24 h, hydrofluoric acid for 3-7 days, hydrochloric acid again for couple of minutes or a few hours. In situ microspores isolated from Thomasites serratus were recovered by dissolving small portions of sporangia with the aid of nitric acid (HNO 3 , 60%) for 24-40 h and potassium hydroxide (KOH, 10%). Palynological slides using glycerin jelly for light microscopy observations were made, and some samples were coated with gold for observation using scanning electron microscopy (SEM). The preparation of samples for the study using transmission electron microscopy (TEM) was more complicated and was also described in the papers of cited authors.
Specimens of the Paleozoic selaginellalean species Thomasites serratus (specimens Nos. The methods for chemical and mechanical preparation of spores for TEM are mentioned in the cited papers but usually involved a few weeks of fixation within a 4% paraformaldehyde solution in a phosphate-sodium buffer; the specimens were embedded in agar-agar, washed in distilled water, and then postfixed in a 1% osmium tetroxide solution in phosphate-sodium buffer for 24 h. The samples were dehydrated in a graded ethanol series for 48 h, then immersed in pure propylene oxide and then in mixtures of propylene oxide and an increasing percentage of Epon resin for 24 h. After transfer to pure Epon resin for 24 h, the samples were embedded in fresh Epon resin using flat molds and placed in a 607 • C oven for 48 h for polymerization. The blocks of resin were stored in a closed box containing silica gel. The resin blocks were trimmed and sectioned with a Reichert Ultracut S ultramicrotome using a diamond knife. Sections were collected on uncoated 300-mesh copper grids and stained with a methanol solution of 7% uranyl acetate for 15 min and an aqueous lead citrate solution for 20 min.

Apical Papillae/Laminated Zones (TAP/LZ)
The sculpture of the proximal surfaces of spores is variable. It is usually laevigate or variously sculptured (granulate, verrucate, spinate, reticulate). Proximal sculpture elements can be of different types [15], sizes, positions and numbers. Very rarely, some spore taxa possess special proximal sculpture elements called three apical papillae (TAP) [16][17][18][19][20]. TAP are usually three verrucae/granae close to the proximal pole among two rays of the trilete mark. Sometimes they are positive sculpture elements, i.e., elevated above the proximal surface of the spore, and can sometimes be observed only in LM or TEM, i.e., they need not occur on the exine surface but among the exine layers. TAP when studied using TEM are called laminated zones (LZs), i.e., it is a proximal structural change of the exospore [18][19][20]. Proximal exospores include three special zones where the inner layer is markedly thickened and tangentially cleft in ten or so lamnae that are irregularly segmented and interlinked, while the outer layer shows some transversal fissures ( Figure 1). These LZs are situated between the aperture arms, near the proximal pole. TAP occur very rarely and are reported from only a few spore-producing genera. TAP/LZs are criteria for the reliable recognition of spores of Paleozoic, Mesozoic and Cenozoic plants of the isoetalean clade [18][19][20][21][22][23][24].
The last Silurian-Devonian TAP miospore is Brochotriletes tripapillatus reported from the Givetian of Libya [36] that differs from all previously mentioned species due to the reticulate sculpture of its distal surface.
Morphologically different Devonian-Carboniferous miospores with TAP are trilete monopseudosaccate species of the genus Endosporites where the central body of these spores sometimes possesses TAP [20,75] that can be seen especially if the central body is separated. Globally, the oldest record of in situ Endosporites with TAP (Figure 2c,d) is from the upper Devonian (Famennian) of the USA, [20] but the genus is typical for Carboniferous strata. Carboniferous TAP spores are very rarely reported when compared to those of Silurian/Devonian age. Although trilete cingulate miospores of the genus Lycospora belong to the most abundant Paleozoic spores with more than fifty species [76], only one, Lycospora tripapillata from the middle Pennsylvanian of USA [77], possesses TAP. Similarly, within another trilete cingulate genus Densosporites (one of the most abundant Carboniferous miospores), only one species, D. tripapillatus from the Mississippian of the USA, has Carboniferous TAP spores are very rarely reported when compared to those of Silurian/Devonian age. Although trilete cingulate miospores of the genus Lycospora belong to the most abundant Paleozoic spores with more than fifty species [76], only one, Lycospora tripapillata from the middle Pennsylvanian of USA [77], possesses TAP. Similarly, within another trilete cingulate genus Densosporites (one of the most abundant Carboniferous miospores), only one species, D. tripapillatus from the Mississippian of the USA, has TAP. The dispersed trilete cingulate miospore genus Crassispora has more than 40 species [78] and its first taxa appeared in the Devonian [79,80]. The most abundant species of the genus is C. kosankei that is defined without TAP [81-83] but sometimes it is described and illustrated with them [77]. Another species, C. plicata, is clearly defined and illustrated only with prominent TAP [84]. The Pennsylvanian miospore genus Thomasospora with the only species T. gigantea (Figure 2a,b,e,f) is also characterized by a distinctive TAP [85], trilete character and equatorial cingulum.
Importantly, we note that all Paleozoic TAP spores are trilete and, except for the monopseudosaccate Endosporites, all of them possess equatorial crassitudo/cingulum and have central/inner bodies that bear trilete marks.

Botanical Affinity of TAP Spores
It is possible to determine the botanical affinity of spores and pollen only by their research in situ, i.e., those isolated directly from the reproductive organs of plants. Stratigraphically, the oldest in situ TAP spores are trilete equatorially crassitate isospores of the Synorisporites papillatus and Aneurospora types isolated from Cooksonia pertonii subsp. apiculispora from the Lochkovian of UK [11]. TAP spores are known only in few species of the trilete crassitate genera Ambitisporites, Aneurospora and Retusotriletes, which are produced by rhyniophytoid plants. Another spore genus with TAP is Brochotriletes but we still do not know its parent plant.
Another Upper Devonian lycopsid Cymastrobus irvingii from the Famennian of Australia [2] yielded trilete in situ microspores with prominent TAP on some probable central bodies but a hypothetical pseudosaccus was not preserved and showed trilete megaspores of the Valvisisporites auritus type (without TAP). Kosoviella timanica from the upper Devonian of Russia [2] produced TAP trilete cinguate microspores, probably of the Cristatisporites type.
Four species of isoetalean lycopsid Leclercqia were palynologically studied including L. andrewsii from the Emsian of Canada, L. complexa also from the Emsian of Canada, L. uncinata from the middle Devonian of China and Leclercqia sp. from the Emsian of Canada [7]. Although spores from all palynologically studied Leclercqia species were compared to the same miospore species, Acinosporites lindlarensis, they exhibited a large range of morphological variation. In situ Leclercqia spores were compared [7] with dispersed Acinosporites lindlarensis Morphon of the late Emsian to the early Givetian. The in situ spores exhibited "palingenesis"; immature spores of L. complexa resemble the mature spores of L. sp. [7], thus connecting the two known plant/spore forms in the Leclercqia lineage. The Leclercqia palynodemes exhibited two tendencies with time: (a) curvatural spinae became larger and (b) the proportion of spores with small distal sculptures decreased.
The taxonomic position of the ligulate lycopsid Leclercqia is interesting. It is the oldest ligulate lycopsid with a strange mix of characteristics: forking leaves, exarch protostele, pitting of tracheids and presence of ligule. It seems that in Leclercqia, various characteristics have been selected by evolutionary pressures at different rates. Leclercqia is usually referred to as a Protolepidodendrale, but sometimes [1] is it stated that Leclercqia is an intermediate between Lepidodendrales and Protolepidodendrales. Based on palynological data, i.e., in situ spores isolated from four species of Leclercqia, we prefer to assign it as a Lepidodendrale than a Protolepidodendrale.
Another isoetalean lycopsid Oxroadia gracilis from the Tournaisisan of UK [2] yielded TAP microspores, probably of the Anapiculatisporites type, and TAP megaspores belonging to the dispersed megaspore genus Setispora. The taxonomic position of Oxroadia gracilis is still not quite clear, and it is usually assigned as transition between Protolepidodendrales and Lepidodendrales but based on the occurrence of TAP micro-and megaspores, Oxroadia belongs to the Lepidodendralean lycopsids.
Taxonomically and phylogenetically important is the lycopsid isoetalean species Bisporangiostrobus harissii from the late Devonian (Famennian) of the USA. Its in situ microspores of the Geminospora lemurata type possess prominent TAP; Bisporangiostrobus belong to the order Bisporangiostrobaceae [86] and is not a member of rhyniophytes but it belongs to a new TAP isoetalean lineage probably evolved from rhyniophytes within the Emsian. Bisporangiostrobus harissii with Geminospora lemurata can be the first member of the nonrhyniophytoid TAP isoetalean lineage because all other Geminospora lemurata producers appeared later [86] and the first dispersed G. lemurata spores appeared within the Emsian [84].
All other Paleozoic in situ TAP spores are from Carboniferous and Permian strata including Endosporites microspores isolated from the Stephanian sub-arborescent lycopsid Polysporia radvanicensis [75] and P. doubingeri [87] and Devonian Polysporia sp. [20]. It is important that Famennian Polysporia sp. from Ohio, USA [20] yielded not only TAP Endosporites microspores but also TAP trilete in situ Valvisisporites megaspores that were produced by several species of Polysporia/Chaloneria [20]. Sometimes TAP in situ microspores isolated from upper Devonian isoetalean lycopsid Cymastrobus irvingi [21] are interpreted as the Endosporites type but potential monopseudosaccus was not observed. In situ trilete cingulate TAP spores of the Crassispora type were produced by the arborescent lycopsids Sigillaria and Mazocarpon (probably a coal-ball counterpart to Sigillaria) [88].
TAP cingulate spores were also produced by some herbaceous lycopsids, e.g., Thomasites serratus from the upper Duckmantian of the Czech Republic that yielded microspores of the Thomasospora gigantea type [85].
Cingulate Densosporites spores with TAP were isolated from the herbaceous lycopsid Porostrobus zeilerii [89,90] from the Mississippian of Spitzbergen, Norway, and the subarborescent species Omphalophloios feistmanteli [78] from the upper Duckmantian of the Czech Republic.
TAP miospore species of the cingulate genus Lycospora has never been macerated from reproductive organs of parent plants, but we can be sure that Lycospora was produced only by arborescent lycopsids of the Lepidodendron type that produced cones of the Lepidostrobus type because we do not know another its producer and the Densosporites miospores are produced by sub-arborescent lycopsids of the Omphalophloios and herbaceous lycopsids of the Selaginella type [7].
Another in situ record [24] is the trilete cavate Permian miospores species Densoisporites polaznaensis with prominent TAP and the associated isoetalean lycopsid genus Viatcheslavia vorcutensis from the Guadalupian of Russia. Three species of the Mesozoic isoetalean lycopsid plant genus Pleuromeia, P. sp. from the Triassic of Romania, P. rossica from the Triassic of Russia and P. sternbergii from the Triassic of Germany [18,19] were palynologically studied for in situ spores. Pleuromeia sp. yielded TAP megaspores of the Talchirella daciae type [18], P. sternbergii produced TAP microspores of the Densoisporites nejburgii type and TAP megaspores Trileites polonicus [18,19] and P. rossica produced in situ microspores assigned to Densoisporites neuburgae from the Mesozoic (Triassic) Pleuromeia rossica and [22] also megaspores of the Trileites polonicus type. TAP are not visible on these megaspore surfaces using LM microscopy but they were clearly seen when looking at the central body using SEM. Interestingly there are not only three papillae but several papillae (a few hundreds) along the rays of the trilete mark; TEM analysis of these megaspores shows the occurrence of LZs.
TAP spores continued through to the Mesozoic and Cenozoic. The occurrence of several isoetalean taxa from the Mesozoic and Cenozoic has been reported [91], sometimes with in situ spores but without a description of TAP spores probably because TEM was not used. The Mesozoic taxa include species such as Skilliostrobus australis, Annalepis latiloba, Cylostrobus ornatus, Isoetes circularis, Nathorstiana arborea, and Isoetites insignis; Cenozoic species include Isoetes reticulata.
It is important that not only microspores but also some in situ megaspores have laminated zones. Several sculpture elements along the rays of the trilete mark of in situ megaspores of the Trileites polonensis type were isolated [23] from the Triassic isoetalean lycopsid Pleuromeia sternbergii. These sculpture elements are laminated zones that were recognised by TEM. The same pattern is possessed by Carboniferous microspores of the Anapiculatisporites type that was produced by Carinostrobus foresmani [90], a member of the selaginellalean clade and probably by [21] Oxroadia gracilis.
Not only mio-and microspores possess TAP; we also have some records of TAP in megaspores. The Mesozoic isoetalean lycopsid Lepacyclotes (Annalepis) zeilerii (Triassic of France [23] yielded TAP trilete megaspores of the Tenellisporites marcinkiewiczae and monolete microspores of the Aratrisporites saturni type. It is interesting that these in situ microspores are not trilete but monolete and possess a central body and monopseudosaccuslike trilete isoetalean microspores of the Endosporites type.
Other in situ TAP isoetalean megaspores are known from Pleuromeia sp. [21,92] that can be assigned to the dispersed megaspore species Talchirella daciae and Pleuromeia sternbergii with TAP megaspores of the Trileites polonicus type.

New Approach to Phylogeny of Paleozoic Isoetalean Lycopsids Based on Palynological Evidence
The generally accepted scheme of phylogeny of Paleozoic isoetalean lycopsids [1] is shown in Figure 3. It suggests that within the Ludlow-Přídolí interval, two main phylogenetic lineages evolved: rhyniophytoids and lycopodialean plants. From the lycopodialean lineage a few lycopsid groups originated including zosterophylls (from the Lochkovian to Givetian), asteroxylaleans (from the Lochkovian to Emsian) and drepanophycaleans (from the Pragian to Famennian). An independent main lycopodialean lineage continues from the Přídolí to today.  Figure 3. Phylogeny of isoetalean lycopsids, modified from [1].  It was supposed [1] that two significant lineages, selaginellalean and isoetalean, originated from the group of protolepidodendralean lycopsids. However, the palynological problem of this scheme is that TAP spores occurring in some rhyniophytoids were never been recorded in any member of the protolepidodendraleans., i.e., TAP spores would have had to disappear ("hiatus") and appear again in spores of the isoetalean clade some tens of millions of years later. This seems less probable.
It is possible to postulate a new hypothesis supported by the palynological data and by the uninterrupted continuation of a TAP line from the lower Silurian (Wenlockian) to Cenozoic. If TAP spores are included into this scheme, we can see a different hypothetical phylogenetic scheme of Paleozoic lycopsids, especially the isoetalean and selaginellalen clades (Figure 4). Protolepidodendrales could evolve from rhyniophytoids within the Lochkovian as a "blind" line disappearing in the Tournaisian (Mississippian, i.e., lower Carboniferous) and not in the Frasnian (Devonian) as assumed [1]. The TAP line could originate in the Wenlockian within rhyniophytes because the first TAP spores (Ambitisporites triapillatus) were described here for the first time. TAP spores are reported continuously from the Wenlockian to Givetian within rhyniophytes and from the Devonian (Emsian), Carboniferous and continuing to the Mesozoic and Cenozoic.   Some spores of plants of the selaginellalean clade also possess TAP, i.e., they probably did not evolve from protolepidodendraleans (that lack TAP spores) as believed [1] but from a TAP-independent lineage probably within the middle/late Devonian.

Conclusions
We can understand isoetalean lycopsids producing TAP spores as an independent phylogenetic lineage. The principle of the TAP hypothesis is that this lineage continued from the Wenlockian (lower Silurian) through the Devonian, Carboniferous, Permian, Mesozoic and Cenozoic as documented by dispersed and in situ TAP spores. The TAP lycopsid lineage consists of the selaginellalean and isoetalean (isoetales and lepidodendrales) clades that did not evolve from protolepidodendraleans but directly from rhyniophytoid plants. This hypothetical concept reflects the palynological data that were not included into previous phylogenetic schemes. Based on the palynological data, we assigned the lycopsid genera Leclerquia and Oxrodia to Lepidodendrales and not to Protolepidodendralean plants.
It seems that Paleozoic TAP isoetalean microspores should be divided into two groups. The first group consists of Silurian-Carboniferous trilete TAP spores of some species of the genera Ambitisporites, Synorisporites, Streelispora, Brochotriletes, Thomasospora, Crassispora, Lycospora, Densosporites and Cristatisporites. The second group are some Devonian-Carboniferous trilete TAP cavate/pseudosaccate species of the genera Acinosporites, Geminospora and Endosporites. Mesozoic isoetalean TAP spores are morphologically similar. TAP/LZs are usually three verrucae/granae close to the proximal pole between the rays of the trilete mark but their number can be higher (some tens or even a few hundreds).