Secondary phloem in arborescent lycopsids

Abstract

Consensus regarding arborescent lycopsid gross anatomy has not been reached in 150 years of study (Williamson, 1872; Walton, 1935; Pannell, 1942; Felix, 1952; Andrews & Murdy, 1958; Eggert, 1961; Bateman, 1994; Krings et al., 2011; Boyce & DiMichele, 2016; Hetherington et al., 2016; D'Antonio & Boyce, 2020; D'Antonio et al., 2020; DiMichele & Bateman, 2020). By comparison, the idea that they did not produce secondary phloem originated in the mid-19th century (Williamson, 1872). The nature of the cambium in arborescent lycopsids and whether or not secondary phloem might have been produced were debated in the early literature, often in association with unsupported speculation regarding unique cell types and wall compositions (Scott, 1900; Seward & Hill, 1900; Weiss, 1901; Seward, 1902). However, after this early phase of study, expectations coalesced on a lack of secondary phloem as has been widely accepted ever since (Seward & Hill, 1900; Seward, 1902; Arnold, 1960; Eggert & Kanemoto, 1977; Stewart & Rothwell, 1993; Taylor et al., 2009). Arborescent lycopsids produced a paucity of primary phloem immediately peripheral to the primary xylem (Eggert & Kanemoto, 1977), in between which secondary xylem would be intercalated by cambial activity (Cichan, 1985). As wood production progressed, the primary phloem would be displaced further from the center of the axis, resulting in both radial compression against more peripheral cortical tissues and tangential stretching as the circumference that the phloem must occupy increased with the radius. This expansion in circumference with continuing wood production was accommodated in the cambium itself primarily by an increase in the cell diameter of each cambial initial, resulting in larger and larger wood cells (Cichan, 1985). At a minimum, however, such growth would have disaggregated the primary phloem into separate strands and might reasonably be expected to have terminated its activity entirely. Thus, the long-held understanding that secondary phloem was absent from arborescent lycopsids presents a problem in that a reliance exclusively on primary phloem functionality in a woody tree should interfere with continued function over a life span measured in decades (Phillips & DiMichele, 1992) to centuries (Boyce & DiMichele, 2016). The studied specimen of a Stigmaria ficoides rooting system (Coal Ball 532, Peel B top, Field Museum (Chicago, IL, USA)) comes from the Herrin Coal (Sahara Mine No. 6), Middle Pennsylvanian (315–307 Ma) of North America, a major coal ball deposit (e.g. DiMichele & Phillips, 1994). Only one peel was available with the stele in transverse section; longitudinal sections of the extraxylary tissue were not present in the collection and additional sectioning was not possible because the source coal ball would have been discarded decades ago. The specimen was noticed as possessing unusually well preserved extraxylary tissue during a survey of the lycopsid steles preserved in acetate peels derived from c. 250 Euramerican calcium carbonate ‘coal ball’ concretions for an unrelated project. Given the poor preservation potential of phloem in general (Smoot & Taylor, 1978), any specimen ostensibly exhibiting phloem fossilization – and particularly fossilization of secondary phloem – warrants particular attention. Micrographs of the specimen were digitally analyzed with ImageJ software (NIH, Bethesda, MD, USA). Measurements and descriptions of potential fossil phloem were compared against both those of ray tracheids from the same specimen and those of phloem in Isoetes tuckermanii measured from fig. 49 of Karrfalt & Eggert (1977). Data are presented in Table 1. The transverse stele can be identified as S. ficoides based upon the centrifugal increase in wood tracheid diameter, the division of wood into wedges, and conspicuous lateral appendage traces in the wood (Frankenberg & Eggert, 1969; Cichan, 1985) (Fig. 1a,b). The major and minor axes of the stele were 9.2 and 6.3 cm; wood thickness was 3.2 cm. A thin layer of tissue (maximum thickness = 800 μm) peripheral to but anatomically contiguous with the wood cylinder is present along the stele's entire circumference (Fig. 1b). This tissue is composed of unordered cells polygonal to ellipsoidal in transverse section with lumen diameters ranging from 15 to 65 μm and cell wall thickness of c. 5 μm – smaller and with thinner walls than the adjacent wood tracheids of 60–250 μm diameter with walls c. 8 μm thick. Within the wood itself are scattered discrete patches of low-diameter tracheids similar to what has been termed interrupted wood or a disruptive zone (Frankenberg & Eggert, 1969; Cichan, 1985) (Fig. 1c). The stigmarian extraxylary tissue and interrupted wood have overlapping ranges of lumen diameters (15–65 vs 13–90 μm) and similar cell wall thicknesses, but they differ in location and in that the extraxylary tissue encompasses the entire circumference of the wood cylinder. The extraxylary tissue is identified as secondary because wood rays, including ray tracheids, continue out from the wood through the full thickness of the tissue (Figs 2, 3a). As the rays are cambially derived, tissues internal to the outer edge of the wood rays are interpreted as being produced by the cambium as well and therefore as being secondary in ontogeny. Importantly, a cell type is present within this secondary tissue external to the xylem that is not seen elsewhere in the specimen. Cells of this type possess bar-like thickenings that traverse the cell lumen in the view of the peel (Figs 2, 3a). These features appear to be end walls of cells following the long axis of the stele rather than lateral walls of cells oriented radially. Tracheids normally taper, so that a section near their terminus can be expected to be oblique to the long axis of the stele, whereas the end walls seen here are flatly transverse – more like phloem. Furthermore, the thickenings and spaces between thickenings are of dimensions that are distinct from those of the ray trachieds (see Table 1 for a full comparison). Given their placement and orientation, these cells can be neither ray tracheids nor transfusion tracheids. Rather, they are comparable to sieve plates in the phloem of stigmarian appendages (Stewart, 1947). Moreover, their cellular location, orientation, shape, size and histology are consistent with those of sieve cells in the phloem of extant lycopsids including Isoetes – the closest living relative to the arborescent lycopsids (Karrfalt & Eggert, 1977) – and they are also like those of Chaloneria cormosa, a related rhizomorphic lycopsid (Pigg & Rothwell, 1983; Bateman et al., 1992). Given the information available, the most reasonable interpretation of this tissue – unambiguously secondary in origin – is as phloem and, thus, as secondary phloem. A vascular cambium cannot be positively identified in this Stigmaria specimen (or in previously described material: Seward, 1902; Frankenberg & Eggert, 1969). To that end, the fossil appears to be in a state representing the termination of secondary growth, consistent with the limited and determinate growth of lycopsid wood (Eggert, 1961; Cichan, 1985). With rare exception among sphenophytes (Eggert & Gaunt, 1973; Wilson & Eggert, 1974), secondary phloem is typically associated with the lignophyte clade that includes the seed plants and their extinct progymnosperm ancestors (Meyer-Berthaud et al., 2010; Spicer & Groover, 2010), where it is the peripheral product of a bifacial cambium (Esau, 1953; Beck, 1960). We cannot rule out the possibility of an entirely different process here. Paleozoic lycopsids had the capacity for bifacial secondary growth: their secondary cortical tissues (i.e. periderm) could involve production of tissue both to the periphery and towards the center relative to their cork cambium or phellogen (e.g. D'Antonio & Boyce, 2020). With no homology between secondary growth in lycopsids and seed plants, however, the possibility of alternative mechanisms of secondary tissue production should be considered. If a bifacial cambium existed between the wood and more peripheral secondary tissue, that cambium has entirely differentiated here. Alternatively, the terminal differentiation of a unifacial cambium could have led to a transition to the production of other cell types, including phloem. These alternatives might be expected to have different implications for the phloem network throughout the plant. Additional material will be needed to assess the developmental potential of the arborescent lycopsid cambium. Regardless of bifacial vs unifacial activity, other aspects of lycopsid cambial ontogeny are crucial for interpretation of the extraxylary anatomy. A standard expectation for secondary tissue is orderly radial cell files, and their absence may have been a barrier to the recognition of lycopsid secondary phloem in the past. In any woody plant, the circumference represented by the cambium must increase along with the radius of the wood cylinder as new secondary xylem cells are added centripetally. In arborescent lycopsids, this increase was mainly accomplished by increasing the size of initials, resulting in wider tracheids, rather than by anticlinal divisions of those initials that would initiate new cell files as in most seed plants (Cichan, 1985). However, functional phloem cells are much smaller than stigmarian wood tracheids (Eggert & Kanemoto, 1977). Periclinal cambial divisions would add cells to the secondary phloem in discrete files, but additional divisions would have been a subsequent necessity as part of differentiation to arrive at appropriate diameters of functional phloem cells. This process of phloem differentiation is proposed to have obscured the original cell files that otherwise would be present as seen in the adjacent secondary xylem (additionally, some of the increase in circumference was accommodated by dilation of the rays as they traverse the secondary phloem (Figs 2, 3), which is consistent with some living dicot angiosperms (e.g. Lev-Yadun, 1996)). The distinctiveness of this extraxylary tissue has sometimes been recognized with descriptors such as ‘terminal wood’ (Frankenberg & Eggert, 1969); these tissues may well be secondary and specifically phloem. With this search image in place – ray extensions beyond the wood delineating a secondary tissue despite lacking clear cell files – extraxylary tissue conforming in organography to that in this Stigmaria specimen can be recognized as having been illustrated before in Lepidophloios (fig. 1 in Seward (1902), where it was identified as Lepidodendron) (Fig. 3b), indicating both that this is indeed a more general phenomenon, not merely a happenstance in our specimen, and that it extends to the aerial shoot and not just the rooting system. The tissue in that specimen was correctly recognized as secondary but was interpreted as composed of small tracheids and cambial initials (Seward, 1902). However, this interpretation could not be correct given that the large size differential between these purported initials and the mature tracheids would have only exacerbated stresses associated with the increasing circumference of the cambium as new daughter tracheids expanded and matured. Rather, the similarities between that tissue previously described from Lepidophloios and the Stigmaria specimen described here (i.e. relative sizes of tracheids and extraxylary tissue cells, relative thickness of the extraxylary tissue, continuation of wood rays out past the wood and into the extraxylary tissue) suggest that the extraxylary tissue in that Lepidophloios specimen is also likely to be secondary phloem, although no potential sieve cells were described or figured. Validation of the search image proposed here for secondary phloem can be sought by looking outside of the lycopsids to analogous systems. Most tracheid-based wood among seed plants involves thin tracheids allowing for continuity of cell files between secondary xylem and phloem with minimal distortion. However, fossil medullosan seed plants had tracheids that could be several hundred microns wide (Smoot, 1984), comparable to those of arborescent lycopsids. Such dimensions are inappropriate for phloem and could be expected to require the disruption of clear cell files because of the many additional cell divisions needed to arrive at smaller diameter cells. Indeed, that expectation is borne out (Smoot, 1984) with the secondary phloem of medullosans scarcely recognizable as secondary except by the extension of the rays out past the wood, exactly as described here in Stigmaria. The secondary vasculature of the extant lycopsid Isoetes (e.g. Foster & Gifford, 1959; Paolillo, 1963) may be another point of comparison for the fossil tissues interpreted here as secondary phloem. The composition of the tissue produced to the inside of the cambium in living Isoetes has been debated – whether it consists of secondary tracheids, secondary sieve cells, parenchyma, or some combination of these cell types (and this may well vary between species; see Paolillo (1963) for a review), as well as whether it can be considered part of the stele at all, given that some interpretations place the isoetalean cambium external to the primary phloem in the primary cortex, and therefore characterize it as extrastelar (e.g. Lang, 1915). The transverse walls of secondary tracheids (plate XII, fig. D in Paolillo, 1963; figs 203 and 207 in Ogura, 1972) are similar to those described from the stigmarian sieve cells here. The uncertainty of secondary vasculature anatomy and uniqueness of secondary vasculature development in Isoetes, however, may limit the utility in comparing this tissue to arborescent lycopsid secondary phloem. For fossil-based discovery, validation always comes with replication across further specimens. Here, the need for additional material is compounded by the secondary tissue of Stigmaria interpreted as phloem coming from a single coal ball represented by just one peel and one line drawing of a thin section over 100 years old. The stigmarian stele is presented in transverse section and additional material, including primary coal balls or slabs, would be needed to prepare longitudinal sections for accessing lateral wall histology of the potential secondary phloem. With such material, direct comparison could be made with the elliptical sieve areas described from the lateral walls of primary phloem of Lepidodendron (Eggert & Kanemoto, 1977) and those on the lateral walls of phloem in stigmarian appendages (Stewart, 1947). It is hoped that our work will provide a search image for further discoveries. In the context of this search, the observation that the potential secondary phloem is specifically in a rooting system may be informative. Phloem is among the most delicate and rarely preserved of plant tissues and roots are overrepresented and better preserved in coal floras relative to stems, because they were already growing within the peat substrate where anoxia associated with waterlogging could inhibit decay (Nelsen et al., 2016). Thus, in the search for additional material, rooting systems grown in the substrate may be more likely sources than aerial shoot material deposited on the substrate, often post-mortem. The presence of secondary phloem in arborescent lycopsids, and the recognition of the importance of roots in the search for additional occurrences of the tissue, may matter for reconstructing the paleobiology of other extinct nonlignophyte plants similarly thought not to have produced secondary phloem. For example, wood production in Rhacophyton, Amoricaphyton and some cladoxylopsids was so scant that an absence of secondary phloem is plausible (Matten, 1974; Meyer-Berthaud et al., 2004; Strullu-Derrien et al., 2014), but wood production could be massive in the Calamites trees that were relatives of the modern Equisetum, and they may have possessed a bifacial cambium in Astromyelon, their root system (Wilson & Eggert, 1974). We thank Andrew Leslie, Andrés Baresch and Bill DiMichele for helpful discussion, Simcha Lev-Yadun, Richard Bateman, Sandy Hetherington and one anonymous reviewer for thoughtful reviews that served to improve the manuscript, Richard Bateman for recognizing that what was considered Lepidodendron in Seward (1902) has since been transferred to Lepidophloios, and Paul Mayer (Field Museum) for access to the Stigmaria specimen. MPD'A conceived of the study, designed the study, acquired specimens, acquired specimen photographs, collected and analyzed data and drafted the manuscript; CKB helped to design the study, coordinated the study, acquired specimen photographs and helped draft the manuscript. All data are present in Table 1.

Topics & Concepts