Phylogenetic diversification of Equisetum (Equisetales) as inferred from Lower Cretaceous species of British Columbia, Canada^†

Subdivision

Euphyllophytina sensu 28

Family

Equisetaceae sensu 12

Genus

Equisetum L.

Species

Equisetum haukeanum Stanich, Rothwell et Stockey, sp. nov., Figs. 1–9 .

Species

Equisetum vancouverense Stanich, Rothwell et Stockey, sp. nov., Figs. 10–14 .

General features

All three anatomically preserved morphotypes of shoots, including Equisetum haukeanum (Figs. 1–9), E. vancouverense (Figs. 10–14), and the underground rhizome (Figs. 15–17) display a combination of characters that fall within the range of variation for comparable structures of living Equisetum species. Shoot segments consist of one or two nodes and adjacent internodal regions (Figs. 1–3, 10, 11). The fossils are comparable in size to the smallest living species of Equisetum (e.g., E. scirpoides Michaux, E. variegatum Schleicher ex F. Weber & D. Mohr; 21, 22), ranging 0.7–2.0 mm in diameter. They are characterized by fluted margins (Figs. 3, 10, 11, 15), that in cross sections represent ridges and furrows. At the internodes, all stems have a hollow pith (Figs. 1, 10–12), which is surrounded by a ring of carinal canals that occur on the same radii as the ridges (Figs. 3, 4, 10–12) and a cortex that contains vallecular canals on the same radii as the furrows (Figs. 3, 4, 12). As in living species of Equisetum, only a small number of metaxylem tracheids accompany each carinal canal at the internodes (Fig. 4), and the positions of ridges and furrows alternate from internode to internode (Figs. 3, 10, 11). The cortex is bounded by a distinct epidermis (Figs. 3, 10, 12 at arrow), within which cortical sclerenchyma is sporadically preserved (Figs. 2, 6, 13 at upper right, 15).

The fossil stems display nodal diaphragms (Fig. 1) where the pith is parenchymatous (Fig. 13), and the carinal canals close to form solid cauline bundles (Fig. 13) that may be fused into a ring of tracheids (Fig. 2). As is also characteristic of living species of Equisetum (11), tracheids of the stele typically have helical or scalariform pitting patterns (Fig. 5, at arrow), but those in the nodal ring display multiseriate pitting (Figs. 5, 16). All the stems branch irregularly (e.g., Figs. 1, 2, 10, 15), with one or two branches diverging from the nodes of aerial shoots and three diverging from one node of the underground rhizome. Not all branches are evident in a single section of most nodes (e.g., Figs. 1, 2, 15). Immediately distal to the level of divergence from the stem stele, the stele of each branch consists of a ring of tracheids that surrounds a parenchymatous pith (Fig. 17).

A prominent leaf sheath, consisting of a ring of leaves connected by thin laminar tissue, occurs at each node (Figs. 3, 10, 11), with the thin commissure extending almost to the apex of each sheath (Figs. 3, 14). Most leaves terminate at the level where they separate from each other (e.g., just distal to the level in Fig. 14). It is unclear whether this is the result of leaf tip abscission or incomplete preservation. In cross sections, each leaf displays a single terete vascular bundle (arrowheads in Figs. 3, 14), that is surrounded by parenchymatous mesophyll. Sclerenchyma also occurs to varying degrees toward the abaxial side of each leaf (Figs. 2, 12 at top, 14).

Aerial shoots are distinguished from underground rhizomes by several features. Whereas the cortical sclerenchyma of aerial stems is restricted to the ridges (Figs. 2, 6, at “s”), underground rhizomes display a continuous cylinder of sclerenchyma (Fig. 15). The leaf sheath of underground rhizomes is typically shorter and less intact than on aerial shoots, and the fossil rhizomes may have a larger number of branches at the nodes (Fig. 15).

Equisetum haukeanum

Aerial shoots of E. haukeanum are identified by five convex ridges, each having a sclerenchyma strand that is triangular in cross section (Figs. 2, 6). Likewise, there are five carinal canals, five vallecular canals, and five leaves that make up each stem segment (Figs. 3, 4). Stems measure 1.2–1.6 mm in diameter at the nodes and 0.7–0.9 mm in the internodes. Leaf sheaths are 3.8–4.5 mm long, with leaves that are lens-shaped and connected by a thin commissure in cross sections (Fig. 3). Individual leaves are 0.4–0.6 mm wide and 0.1–0.5 mm thick with a single terete leaf trace (Fig. 3, at arrow points) and a triangular sclerenchyma bundle on the abaxial surface (Fig. 2, at “s”). Sclerenchyma is developed at the base of the sheath but is absent at more distal levels. (cf., Figs. 2 and 3).

Only one complete internode is preserved, measuring 5.7 mm long. Two branches are produced at each of the adjacent nodes. Branches are smaller than the stems upon which they are produced, measuring approximately 0.6 mm wide. All the branches are preserved for only a short distance beyond the cortex of the stem upon which they are borne (e.g., Fig. 2, at right). At the nodes the stem stele forms a cylinder of tracheids 0.2–0.5 mm in diameter (Fig. 2). Tracheids measure 7–18 µm in diameter and display multiseriate pitting (Fig. 5, at right). The stem cortex is not preserved within the epidermis at some levels (Fig. 3). Elsewhere it consists of isodiametric, thin-walled cells, 13–37 µm in diameter, toward the inside (Fig. 1, at top center), and wedge-shaped bundles of elongated sclerenchyma fibers 13–28 µm in diameter at the periphery (Figs. 1, 2, 6).

In cross sections of stem internodes, the carinal canals measure 25–34 µm in diameter and are flanked by two to five metaxylem tracheids (Fig. 4, at arrow points) that measure 5–9 µm in diameter and have scalariform secondary wall thickenings (Fig. 5, at arrow). Epidermal cells are oval in cross section (Fig. 3), slightly elongated in face views (Fig. 7), and have smooth margins. Stomatal complexes occur at the internodes in bands two stomata wide that are located on the flanks of the furrows (Figs. 6, 7). Stomatal complexes are surficial rather than sunken (Figs. 6, 9). As is characteristic of living species of Equisetum, stomatal complexes of E. haukeanum consist of two prominent subsidiary cells that are overlain by the guard cells (Figs. 7–9) (20). Complexes are oval in surface view, measuring 45–52 µm long and 28–39 µm wide.

Equisetum vancouverense

Stems of E. vancouverense are slightly larger than those of E. haukeanum, measuring 1.4–2.0 mm in diameter. They have seven carinal canals 56–102 µm in diameter (Figs. 10–12), tangentially elongated vallecular canals, and angular ridges at the stem periphery (Fig. 10). In contrast with E. haukeanum, which has convex ridges, each ridge of E. vancouverense is biangulate and concave at the center (Figs. 10, 11). At the internodes, the stem cortex is incompletely preserved, but just above the nodes there are vallecular canals adjacent to the stem furrows (Fig. 12). At these levels, the cortex is parenchymatous adjacent to the vallecular canals, and the periphery of the stem is marked by a distinct epidermis (Fig. 12, at arrow). Cortical parenchyma cells measure 13–33 µm in diameter. At the nodes, the vallecular canals close, and the cauline bundles consist of larger numbers of tracheids (Fig. 13). Also at these levels, the pith is parenchymatous, forming a nodal diaphragm, and the cellular cortex is continuous with the stele (Fig. 13). Toward the inside of the cortex, the cells are parenchymatous, and toward the periphery they consist of sclerenchyma fibers, 8–28 µm in diameter (Fig. 13 at upper right). Also near the nodes of E. vancouverense, there is a distinct endodermis that surrounds each stem bundle (Fig. 13, at E).

In cross sections, the leaf sheath of E. vancouverense consists of a ring of angular leaves that are connected by commisures two to five cell layers thick (Figs. 10–12, at arrows). Near the nodes, the leaves have a slightly convex abaxial surface (Fig. 10), whereas more distally they are abaxially flattened or slightly concave (Fig. 11). There is a distinct adaxial epidermis of angular cells that are larger than the mesophyll cells toward the base of the leaf sheath (Figs. 11, 12) and smaller toward the leaf tips (Fig. 14). Near the nodes, the epidermal cells measure 10–43 µm in diameter. Sclerenchyma occurs as a thin layer within the epidermis around the periphery of the leaf sheath (Figs. 10–12). At the most distal levels, leaves separate from each other (Fig. 14) then terminate. They have a terete vascular bundle (Fig. 14, at arrow heads), parenchymatous mesophyll cells 15–37 µm in diameter, and an abaxial band of sclerenchyma (Fig. 14, at “S”).

One branch diverges from the stem at each node of the available E. vancouverense stems (Fig. 10, at top), and branches at adjacent nodes diverge from opposite sides of the stem. Branches measure 0.5–0.9 mm in diameter. Each branch extends outward through the leaf sheath of the stem (Fig. 10) and displays parenchymatous pith and cortex at the branch base (Fig. 10). More distal levels of the branches are not preserved.

Underground rhizome

Two specimens of underground rhizomes have been recovered. These are 0.9–1.3 mm in diameter, and both show fluted margins that represent ridges and furrows in cross sections (e.g., Fig. 15). However, as in many living species of Equisetum (e.g., E. laevigatum A. Braun.; Fig. 18), the fluting is somewhat less pronounced than in the aerial stems (e.g., Figs. 3, 11). Also, as is characteristic of underground rhizomes in living species of Equisetum (e.g., Fig. 18), sclerotic cortex is more-or-less evenly distributed around the periphery of the underground rhizomes (c.f., Figs. 15, 18). The zone of outer sclerenchymatous cortex is 65–135 µm thick in the fossil rhizomes, and consists of fibers that are 9–34 µm in diameter. Neither rhizome has internal stelar tissues preserved, and in one specimen, the internal cavity is filled with debris (Fig. 15). If the internal anatomy of the fossil rhizomes was as aerenchymatous as, and constructed of delicate tissues similar to those of, the rhizomes of living E. laevigatum (Fig. 18), the absence of cells in the interior of the fossils would be easily explained.

Both fossil rhizomes show some evidence of a leaf sheath at the nodes (Fig. 15, at LS), and both branch. One specimen shows three branches, whereas the other has only one. Branch diameters range from 0.6 to 1.0 mm, and steles within the branch bases consist of a cylinder of tracheids that surrounds a parenchymatous pith (Fig. 17). Tracheids in this region are 11–20 µm in diameter and show multiseriate pitting (Fig. 16). Because all branches are broken near the points of attachment, we were unable to determine whether they represent underground rhizomes, tubers, or aerial shoots.

Growth form and plant architecture

Fossils described in this paper include both underground rhizomes and vegetative aerial shoots. We do not know which, if either (or both), of the E. haukeanum and/or E. vancouverense specimens belong(s) to the underground rhizomes from the same locality. Because aerial shoots of living Equisetum, fossil Calamitaceae and extinct Equisetaceae (where known) all arise from underground rhizomes, it is reasonable to assume that the aerial Equisetum shoots from Apple Bay do as well. If correct, then the plants that produced E. haukeanum and E. vancouverense are similar to living species of Equisetum in overall growth form. Aerial shoots of these fossil morphospecies range from 0.7 to 2.0 mm in diameter, which is comparable to the smallest living species of Equisetum (21, 22). Although there is a theoretical possibility that the fossilized aerial shoots could represent ultimate branches of a larger plant, that alternative is unlikely because the underground rhizomes at Apple Bay fall within the same size range as the aerial shoots. Therefore, one or both of the plants that produced E. haukeanum and E. vancouverense were almost certainly as small as the most diminutive of living Equisetum species (21, 22).

A combination of branching patterns and maximum stem diameters of the Equisetum fossils are informative of plant architecture. Many living species of Equisetum subgenus Equisetum have vegetative shoots that branch profusely at each node to produce symmetrical shoot systems with an overall conical shape (e.g., E. arvense L., E. telemateia Ehrh.). Other species, particularly in Equisetum subgenus Hippochaete, have aerial shoots that either are unbranched (e.g., E. variegatum Schleicher ex F. Webber & D. Moore) or are only sparsely branched (e.g., E. scirpoides Michaux). Such species typically produce only one or two branches at those nodes where branches do occur. If the one or two branches preserved at each node of the fossils are characteristic of the plants by which they were produced, both morphospecies of aerial shoots represent species that were sparsely and irregularly branched.

Interspecific relationships of Equisetum species

Classical studies have divided living species of Equisetum into two subgenera, Equisetum L. (E. pretense L., E. telemateia Ehrhart, E. fluviatile L., E. arvense L., E. diffusum D. Don, E. sylvaticum L., and E. palustre Ehrhart) and Hippochaete Milde (E. scirpoides Michaux, E. variegatum Schleicher ex F. Weber, E. giganteum L., E. laevigatum A. Braun, E. myriochaetum Cham. et Schldtl., E. ramosissimum Desfontaines, and E. hyemale L.). Depending upon the source of data evaluated and the methods of systematic analysis employed, a 15th species, E. bogotense Kunth, resolves as sister to the remaining living species (8), as basal within subfamily Hippochaete (8; 16), or else is placed within the subfamily Equisetum (32). Finer resolution of species relationships has been difficult to achieve from morphological characters (21, 22; 32), but analyses of nucleotide sequences have provided greater, if not uniformly consistent, patterns of resolution (8; 16).

Taxonomic placement of E. haukeanum and E. vancouverense among living species is difficult to assess because fertile structures of the fossils have yet to be discovered and because there is a dearth of known informative morphological characters for inferring relationships within the genus Equisetum (21, 22, 24; 8). Stomatal position and branching pattern are among the very small number of morphological characters that appear to be informative for that purpose (32; 24), and those have been employed primarily for placing species in one of the two subgenera (23). Those characters can be recognized from anatomically preserved vegetative shoots.

Stomata are surficial and usually scattered or in bands two or more wide in subgenus Equisetum (i.e., horsetails), while those of Hippochaete (i.e., scouring rushes) are distinctly sunken below the ordinary epidermal cells and are usually arranged in single lines (23). Aerial shoots of subgenus Equisetum branch in regular whorls or are unbranched, while those of subgenus Hippochaete are either irregularly branched or unbranched. A third feature, position of endodermis (i.e., common exterior ring in subgenus Equisetum and as double common or individual endodermis in subgenus Hippochaete; 23) is not fully consistent within the subgenera (21, 22). 32 has identified additional anatomical and fine-structure characters (with scanning electron microscopy) that appear to be of systematic value (e.g., Table 2), but such characters have not yet been employed in phylogenetic analyses.

Equisetum haukeanum and E. vancouverense share with subgenus Hippochaete an irregular pattern of branching (Table 2). Endodermal positioning in E. vancouverense aligns more closely to subgenus Hippochaete, while stomatal features of E. haukeanum conform more closely to those of subgenus Equisetum, but the data are inconclusive and the systematic reliability of these characters remains untested. Nevertheless, the combination of morphological characters displayed by E. haukeanum and E. vancouverense suggests that these species combine characters of the two subgenera (Table 2). These novel combinations of characters with respect to living species of Equisetum may reflect a Mesozoic species richness that has since been reduced by extinction.

Evolution and phylogeny of Equisetum

There is a long and rich fossil record of equisetophytes that extends from the uppermost Devonian to the Recent and that documents excellent transformational series of morphologies and anatomy from which to infer the evolution of plant organs (i.e., leaves and sporangiophores), of stelar architecture, and of the features that characterize crown group Equisetum (4; 41; 45). Fossils contributing to the transformational series are assignable to several families sensu 4, including Archaeocalamitaceae (Upper Devonian to the Lower Permian), Calamitaceae (Late Mississippian to Permian), and Equisetaceae (most common from the Triassic to the Recent; 41; 45). Other authors (e.g., 12) have chosen to place most of the same species in the single order Equisetales or even in the single family Equisetaceae. The latter classification emphasizes close relationships among many equisetophytes and allows us to focus on the relatively small number of structural changes that have led to evolution of the genus Equisetum (Table 1).

Only a few structural characters differentiate the Mississippian to Permian Archaeocalamites from the Late Mississippian to mid-Permian Calamites, from the Lower Permian to Jurassic Neocalamites Halle, from the mid-Triassic to Lower Cretaceous Equisetites Sternberg, and from modern-appearing Equisetum that ranges from the mid-Triassic to the Recent (Table 1). There is a general progression of changes in the characters of these plants through time that includes (1) a overall reduction in plant size, (2) the loss of leaf dichotomies, (3) progressive fusion of leaves forming a sheath, (4) change from spreading leaves or leaf tips to leaves/leaf tips that are pressed against the stem, (5) change from trilete to alete spores, (6) change from no spore elaters or three elaters to four elaters, (7) changes in the presence/absence of bracts within cones, (8) change from ridges and furrows that are aligned (or opposite) from internode to internode, to ridges and furrows that alternate from internode to internode, and (9) the change from woody to herbaceous (i.e., the loss of secondary vascular tissue) growth leading to crown group Equisetum (Table 1; 41; 45).

If a minimum age for each of the pertinent characters of crown group Equisetum can be determined from available fossils (Table 1) and if the most ancient date at which all of the crown group characters are present can be hypothesized as marking a minimum age for Equisetum, then the fossil record can provide solid evidence for the origin and early diversification of the genus (31). Many of the characters that together characterize crown group Equisetum (i.e., stem diameter, leaf shape and orientation to the stem, spore type, alternating positions of carinal canals [or ridges and furrows] from internode to internode; Table 1) are relatively well established from known fossils (e.g., 2; 29; 7), but a few require additional evidence or interpretation. The well-known change from alternating whorls of sporangiophores and bracts in Paleozoic calamitalean cones to only sporangiophore whorls in living Equisetum is not as clear an evolutionary transition as is popularly believed. Cones of Archaeocalamites, which predate Calamites, are similar to modern Equisetum in that they typically produce only sporangiophores (47; 1). Likewise, not all living Equisetum plants have terminal cones that consist of only sporangiophores. Well-known Equisetum teratologies include alternating sporangiophore and branct whorls, alternating fertile and vegetative zones along the stem, and irregularly occurring bract whorls within cones (e.g., 33). Such teratologies and the extinct equisetophytes they resemble are reviewed by 30 .

The transition from trilete to alete spores is incompletely documented because in situ spores have not been characterized for many extinct Mesozoic equisetophyte species (27). However, from the specimens that have been described, the transition appears to have occurred during the Triassic (for a detailed discussion, see 27). For example, about 70% of the spores of the Lower/Middle Triassic Equisetum laterale are alete (14), whereas spores of the Upper Triassic Equisetites arenaceus are either alete or display a small trilete mark (27). Alete spores characterize most Jurassic and more recent species. Similarly, the presence/absence and number of spore elaters is not known for most fossil equisetophyte species (e.g., 48). Three elaters characterize Paleozoic calamitaleans (12; 13). In contrast, elaters are not preserved in Triassic species (but see possible exception for E. arenaceus; 27). The first well-documented occurrence of elaters on spores from Equisetum sporangia is from the Middle Jurassic Equisetum columnare (19; Table 1).

With the discovery of essentially modern Equisetum internal anatomy in E. haukeanum and E. vancouverense, we now have evidence that all of the diagnostic characters of crown group Equisetum species had evolved by the Lower Cretaceous (Table 1). Moreover, because there are at least two Early Cretaceous species of Equisetum present in the same depositional setting at the Apple Bay locality, and because both fall at the lower end of the size range for living Equisetum species, it is highly unlikely that E. haukeanum and E. vancouverense represent the initial diversification of the genus Equisetum.

As is illustrated by the Equisetum morphospecies present at the Apple Bay locality, well-dated extinct species provide a strong minimum age determination for the origin and evolutionary radiation of clades (31) that can be regarded as a test of clade age assessments derived from other methodologies (e.g., 8). There are several levels of rigor with which the available fossil evidence can be employed as tests of clade age-hypotheses. The least rigorous level employs fossils that display features that all are concordant with those of the crown group, but that are only a subset of the synapomorphies that characterize the crown group. The oldest widely recognized evidence of this type for the genus Equisetum is represented by E. laterale Phillips from the mid-Triassic of Queensland, Australia; Gould, 1968). Equisetum laterale is preserved by coalified compression and displays all of the morphological features of living Equisetum species. However, the E. laterale fossils provide little evidence for the characteristic internal anatomy or spores of the genus. If this species does represent crown-group Equisetum, then the genus is at least 230 million years old.

A higher level of rigor requires that minimum clade age be inferred from an assemblage of extinct morphotaxa that together document all of the synapomorphies of living species. Using this criterion, E. haukeanum and E. vancouverense provide both of the remaining characters (e.g., characteristic internal stem anatomy and small plant size; Table 1) needed to confidently set a minimum clade age for crown group Equisetum of Early Cretaceous (i.e., at 136 Ma; Table 1).

The most cautious approach for determining minimum clade ages from the fossil record recognizes the possibility of mosaic character evolution among related species. A minimum age is not acceptable unless all of the diagnostic characters of the crown group are known from a single extinct species. In the case of equisetophytes, there is excellent evidence that Mesozoic diversity was far greater than for the extant flora (45) and therefore that a large percentage of morphotaxa displaying subsets of crown group synapomorphies could possibly represent stem group representatives of the clade. Because extinct species of fossil plants are infrequently complete enough to document the whole suite of synapomorphies for a crown group, that level of rigor for testing hypotheses can rarely be applied. For the genus Equisetum, the fossil record has yet to provide data of this type from a single species.

In this regard, there is one species of fossil equisetophytes that prompts us to proceed with caution. Equisetites lyellii (Mantell) Seward is well represented in lowermost Cretaceous sediments (i.e., Berriasian-Valanginian; 48) and is therefore only slightly older than the morphospecies Equisetum haukeanum and E. vancouvenrense (i.e., Valanginian/Hauterivian boundary). Equisetites lyellii is represented by specimens that display excellent evidence for growth architecture, vegetative morphology, reproductive morphology (i.e., peltate sporangiophores), spores, cuticular features, and some internal anatomy of the sporophyte (Table 1). That species displays all the synapomorphies of crown group Equisetum except for the absence of spore elaters and a larger number of metaxylem tracheids surrounding carinal canals of the stem than is present in living species (Table 1; 48).

Clearly, clade age for the genus Equisetum is difficult to establish with certainty from either questionable molecular clock calculations (i.e., unknown levels of nucleotide substitution rate heterogeneity) based on systematic relationships of contested accuracy (i.e., euphyllophytes; 37) that are inferred from highly restricted taxon sampling (i.e., living species only; 34; 8), or from incomplete data for extinct species of Equisetum plants. Nevertheless, the fossil record does provide valuable evidence for testing hypotheses of systematics, phylogeny, and clade ages. With the characterization of E. haukeanum and E. vancouverense from the Lower Cretaceous, we now know that all of the synapomorphies of living Equisetum had evolved by at least 136 Ma (Table 1) and therefore that the crown group probably originated during the Mesozoic. Although the mid-Triassic date for potential crown group Equisetum (i.e., E. laterale) remains to be verified by internal anatomical and spore data (Table 1), that species is only one example of a rich and diverse fossil record (41; 45) that is continuously providing new and more convincing evidence that equisetophytes are derived from among trimerophyte-grade species of the Euphyllophytina (sensu 28; 10), that stem/leaf organography has evolved independently among several clades of euphyllophytes (41; 10; 46; Sanders et al., 2009), and that the equisetophyte clade has been distinct since at least the Upper Devonian (12; 1; 36; 45).