2017/12/28 by Barry A. Thomas, Christopher J. Cleal · 1 citation
Earth and Planetary Sciences · Engineering · Agricultural and Biological Sciences · #Coal and Its By-products #Hydrocarbon exploration and reservoir analysis #Plant Diversity and Evolution
paper · pdf · doi:10.1111/nph.14903
openalex publication_date 2017/12/28 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/15
The arborescent lycophytes formed extensive forests in the Pennsylvanian (late Carboniferous) equatorial coal swamps. They can in some ways be regarded as the dinosaurs of plant life, being an extinct group of organisms that often achieved enormous size and had a major impact on the environment. Reconstructions of the plants from their fossilized remains have revealed a diversity of form, with some of them, like Lepidodendron, achieving statures of up to c. 45 m, with trunk diameters up to 2 m (Thomas & Watson, 1976; Thomas & Seyfullah, 2015b). Their growth patterns were determinate, starting as an unbranched leafy stem that grew vertically and expanded in diameter until it finally branched in a variety of ways, producing either a crown of dichotomizing branches or a series of lateral dichotomizing branches (Thomas, 1978; Bateman et al., 1992; DiMichele et al., 2013). Sigillaria, by contrast, did not grow as tall, and the evidence suggests that it only branched a few times at most and in some species not at all (Hirmer, 1927). Especially in early–middle Pennsylvanian times, these trees were largely responsible for producing thick deposits of peat that have since changed into economically important coal (Montañez, 2016). A key issue to understanding this is the speed of growth and life cycle of the plants. Until recently they have been assumed to have been very fast growing (e.g. Phillips & DiMichele, 1992; Cleal & Thomas, 2005) but this has recently been challenged (Boyce & DiMichele, 2016). We will here examine some of the evidence involved. All plants respond to the same environmental pressures by using the same physiological processes; and face the same trade-offs in water-transport morphologies that maximize flow volume but minimize effort in anatomical change (Tyree & Zimmermann, 2002). The arborescent lycophytes achieved this by having a relatively small amount of highly efficient secondary xylem for water conduction (Cichan, 1986), but with support provided by a thick, peripheral secondary cortex, or periderm. This growth pattern is quite different from that of modern conifer and angiosperm seed-plant trees where both water conduction and support are provided by large amounts of secondary xylem. Logan & Thomas (1987) showed the decorticated tissue of Lepidodendron Sternberg and Lepidophloios Sternberg to have three of the four components of lignin (p-hydoxybenzaldehyde, vanillin and acetovanillone) while that from Sigillaria Brongniart had all four components (also with syringic aldehyde). These lignin components would make the periderm tissues water resistant as well as being supportive. The economic construction of these plants with their low proportion of secondary wood suggests that they could grow very rapidly. The discoveries of clusters of stems and of their rooting bases called Stigmaria (McGregor & Walton, 1972; Gastaldo, 1985a,b, 1986; DiMichele & DeMaris, 1987; DiMichele et al., 1996, 2009; Thomas & Seyfullah, 2015b) indicates that such stands of lycophytes grew simultaneously (Fig. 1). Furthermore, it is most likely that, when the trees were in their unbranched early stage of development (Fig. 2), the emerging forests would have been well-illuminated environments. It is only when the taller, maturing plants started to branch that their lower parts would be in shade with a subsequent reduced ability to photosynthesize (Fig. 3). Boyce & DiMichele (2016) make the point that there is an apparent absence of juveniles, which might suggest the plants did not have rapid growth. However, as Boyce & DiMichele (2016) themselves point out, juveniles will almost certainly not be preserved; and it might anyway be difficult to distinguish a juvenile (‘sapling’) in a coal ball from an unbranched segment of terminal crown. Phillips & DiMichele (1992) estimated that their life span was no greater than 10–15 years and Cleal & Thomas (2005) used this figure to estimate the total carbon budget of a lycophyte-dominated Middle Pennsylvanian swamp forest. This was based on combining: (1) volume calculations for an average lycophyte tree; (2) empirical measures of tissue carbon densities (Baker & DiMichele, 1997) of the periderm and wood that made up the bulk of the volume of each tree; and (3) estimates of the density of trees growing within the swamp (e.g. DiMichele et al., 1996, 2001). Using the Phillips & DiMichele (1992) estimated life span of the individual trees, Cleal & Thomas (2005) then calculated that the trees of these coal swamps would be responsible for taking 160–578 t ha−1 yr−1 of carbon from the atmosphere, and from this it was estimated that the waning of the forest area in later Middle Pennsylvanian times would have had a 2–5 ppm yr−1 impact on atmospheric CO2 levels. Boyce & DiMichele (2016) have recently proposed an alternative model for the growth of these arborescent lycophytes, suggesting that the extraordinarily high productivity value resulting from the calculations of Cleal & Thomas (2005) was untenable, even though their calculations were based on the original Phillips & DiMichele's (1992) estimated life span. Boyce & DiMichele (2016) stated that such growth rates would have been nearly 20 times higher than in modern angiosperm-dominated tropical rainforests and almost two orders of magnitude higher than actual living lycopsids (Brodribb et al., 2007). They instead suggested that the lifespan of these trees must have been much longer than the 10–15 years proposed by Phillips & DiMichele (1992), and more in the order of centuries. However, we believe that there are flaws in their reasoning and that the growth of these arborescent lycophytes was relatively fast. Since many of the plants seem to have favoured disturbed habitats (as suggested by Phillips & DiMichele, 1992) it would surely have made little sense for them to have slow growth rates. We know especially from the remains of the so-called clastic swamps (sensu Gastaldo, 1987) that the Lepidodendron plants would often be destroyed by flooding/cyclone events, and the area would then soon after be re-colonized by a subsequent generation of trees. These clastic swamps are too common to be explained away by mere chance. These lycophytes were clearly adapted to this strategy and it would make no sense if they did not grow quickly and reproduce before the habitat was destroyed. Although there may have been some variation in life spans, it seems unlikely that those forms with crowns of branches with terminal cones would have taken a century or more growing before they reproduced. There are no modern analogues for such a strategy. One key problem is that Boyce & DiMichele try to use modern plants to explain these lycophytes, despite Phillips & DiMichele (1992) having clearly explained why this can be misleading. Boyce & DiMichele (2016) believed there to be no living examples of trees showing such fast growth but there are in fact many examples of very fast growing tropical trees, especially in their early years. For example the Molucca Albizia (Falcataria moluccana (Miq.) Barneby & J.W.Grimes) grows > 20 m in 7 years (Bradley 1922) and can quickly colonize disturbed areas in wet environments. Lebrun & Gilbert (1954) gave the early growth rate for the White afara (Terminalia superba Engl. & Diels) as 2.8 m yr−1 and that for the Corkwood trees (Musanga cecropioides R.Br. & Tedlie) as 3.8 m yr−1. Plantations of Balsa (Ochroma lagopus Sw.) have been found to grow at 5.5 m yr−1 (Anonymous, 1960), while a young individual of Cedrela odorata L. in Nigeria grew at an average rate of 6.7 m yr−1 (Longman & Jeník, 1974). There is also the problem when using modern-day plants to interpret the Palaeozoic arborescent lycophytes that they have completely different body plans. Most trees today have a clear demarcation between the photosynthetic foliage, the supportive trunk, and the foundational and absorptive rooting structures. Nearly all of the productivity in such trees is being generated by only a part of the plant – the foliage. In the arborescent lycophytes, by contrast, photosynthetic activity took place on much of the plant surface, including the leaves, the supporting main stem covered with leaf cushions and even parts of the rooting rhizophores (Frankenberg & Eggert, 1969); it has even been suggested that the periderm may have consisted of living cells and could have been at least partly photosynthetic (Phillips & DiMichele, 1992). This must all surely have had a major impact on the productivity of the plants. One of the facts emphasized by Boyce & DiMichele (2016) was that the leaves were weakly vascularized and had few stomata (Graham, 1935), which would have restricted their photosynthetic efficiency. There is often an assumption that the trunks and the larger branches shed their laminae quickly thereby reducing the photosynthetic ability of the arborescent lycophytes. This is clearly not the case as it has been shown that even quite large stems still retained their leaves (Thomas, 1970; Leary & Thomas, 1989). As pointed out by Boyce & DiMichele (2016), leaf laminae abscission left their persistent, taxonomically distinctive, leaf bases (cushions) that persisted until they were sloughed off from the largest/oldest stems by secondary growth leaving an exposed surface of periderm; see Thomas (1970, 1978), Wnuk (1985) and Opluštil (2010) for more details on this. It is also important to realize that the leaf cushions could move apart to some extent by the expansion of the inter-cushion epidermis during the early stages of secondary growth (Thomas, 1966, 1970) allowing the outer photosynthetic surface to remain attached to the stem even after a significant increase in girth. Because of the unusual growth pattern of these lycophytes, the diameter of the branches does not reflect their relative ages. Dichotomizing large apical meristems resulted in large branches that themselves dichotomized. Each dichotomy resulted in smaller and smaller branches with less and less secondary thickening, so possibly it was only in the lower parts of the main trunk that the leaf bases were sloughed off leaving the rest of the plant with permanent leaf cushions, and the most terminal branches with persistent leaf laminae. One major problem in Boyce & DiMichele's (2016) argument is that they appear to have ignored the fact that the surfaces of the leaf cushions of Lepidodendron Sternberg, Lepidophloios Sternberg, Sublepidophloios Sterzel and Ulodendron Lindley and Hutton and the surface of some lycopsid stems, such as Bothrodendron Lindley and Hutton and Sigillaria Brongniart had large numbers of stomata (Thomas, 1966, 1967b, 1968, 1970, 1974, 1977; Thomas & Seyfullah, 2012). The presence of so many stomata on the leaf cushions (Fig. 4), up to 250 per mm2 in Lepidodendron aculeatum Sternberg and 450 per mm2 in Lepidodendron dichotomum Sternberg (Thomas, 1966, 1970), would have made the plants much more photosynthetically active than might be otherwise assumed if only the more ephemeral leaf laminae stomata were taken into account. It is practically impossible to be exact in giving the total number of leaf cushion stomata on a mature large lycophyte. However, taking into account diameters of stems of known species it is possible to estimate the number of stomata in small sections of stem. For example in a 100-mm length of a 380-mm diameter stem of Lepidodendron aculeatum Sternberg (e.g. Leary & Thomas, 1989), using the median value of 225 stomata per mm2 for only 85% of the cushion area to compensate for the leaf scars, external parichnos and edges of the cushions, the total number of stomata in this relatively small section of stem is c. 23 million. Shoots of Lepidodendron dichotomun Sternberg are relatively narrow (c. 50 mm), so in a 100-mm length, using the value of 450 stomata per mm2 for only the lower parts of the cushions (i.e. 50% because there are virtually none on the upper half) the total number of stomata would be c. 3.5 million. These calculations take no account of the numbers of stomata that would have been on the leaf laminae before they were abscised, but in both of these species the number of active stomata on the plant as a whole would have been immense. There are a few studies on the cuticle of such leaf laminae that show they have a thinner cuticle and more widely spaced stomata than the basal, clasping leaf bases (cushions) (Thomas, 1967a, 1968; Thomas & Seyfullah, 2012, 2015a). Beerling (2002) also confirmed that in the persistently leafy Ulodendron majus Lindley and Hutton and Ulodendron landsbergii Thomas the stomatal index of the leaf laminae is 20–30% lower than that of the leaf cushions. The stomatal indices calculated by Beerling (2002), together with stomatal densities from Thomas (1966, 1967a,b, 1968, 1970, 1977), are shown in Table 1. Beerling (2002) made quantitative Carboniferous and Permian atmospheric CO2 estimates using the stomatal characteristics of arborescent lycophytes and commented that if this effect is systematic across all the arborescent lycophytes it would give an underestimation of atmospheric CO2 levels of 7 to 10%. It was suggested that such an increased level of atmospheric CO2 would have induced an increased level of photosynthesis stimulating higher levels of growth in the arborescent lycophytes and indeed in all the plants growing at this time. This may be true, but growth rates could be, in part, limited by the accessibility of nutrients and/or essential elements such as nitrogen, potassium, magnesium and manganese. Green (2010) suggested that the arborescent lycopsids had an unusual metabolism similar to that of modern Isoëtaceae by utilizing a pathway for carbon fixation that involved uptake of sedimentary carbon and enrichment of CO2 in internal gas spaces; a metabolism related to ‘aquatic CAM’. The idea that such a metabolism could generate high growth rates was disputed by Boyce & DiMichele (2016) by arguing that the gases would be unable to defuse 40–50 m along the trunk. However, such an unusual metabolism could have been a major factor in enhancing the initial trunk growth and the growth of its stigmarian base. Then as the plant grew taller and larger the more usual photosynthetic pathway may have become dominant although the stigmarian bases may still have continued to uptake sedimentary carbon to enhance its metabolic rate. Graham (1935) in his anatomical study of the leaves of arborescent lycophytes showed that, although some leaves had limited numbers of stomata in two furrows on their abaxial surfaces, others had large numbers of stomata spread along two broad bands on their abaxial surfaces. Leaves are clearly important because they are the terminal location in the pathway of water from the soil to the atmosphere and the supply of water to the leaves is of critical importance for photosynthesis and plant growth. Much of the water would have been lost through stomata on the leaves and leaf cushions by evaporation. For example, in most temperate angiosperm plants, > 300 molecules of water are evaporated from the leaf for every molecule of CO2 fixed through photosynthesis (Taiz & Zeiger, 2010). Being tropical wetlands, the temperature and humidity would have been consistently high in the Coal Measures swamps so the loss of water from the leaves would have had some cooling effect on the plants. Because stomata are the main interface for gas exchange between the interior of the leaf and the atmosphere, it is generally accepted that any morphological features are adaptations to environmental factors that affect transpiration and photosynthesis. The stomata on most arborescent lycophyte leaf cushions have their guard cells sunken in pits (Fig. 4), which might be thought of as an adaptation to reduce water loss. They do not, however, show the other adaptation often shown by xerophytes in having stomata clustered in depressed epidermal areas called ‘stomatal crypts’ (Napp-Zinn, 1973; Metcalfe & Chalk, 1979). In a study of the gas exchange in the cloud forest species Drimys winteri J.R. Forst. & G. Forst., Feild et al. (1998) showed that there was no evidence that sunken stomata caused a significant increase in resistance to water loss. Jordan et al. (2008) also showed that there was no relationship between stomatal structure and climate in their study of the Proteaceae and that many sunken, or even encrypted, stomatal types did not appear to be drought related. Sunken stomata in the arborescent lycophytes cannot possibly indicate a dry habitat because the plants grew in waterlogged swamps. The other possibility is that pits may facilitate the diffusion of CO2 to the sites of photosynthesis (Roth-Nebelsick et al., 2009), which could be the case in the thick leaf cushions of arborescent lycophytes with densely packed mesophyll cells. If this is the case, it is another argument for the leaf cushions having such an important role in photosynthesis, metabolism and growth of the plants. Boyce & DiMichele (2016) made the point that increased CO2 uptake and photosynthesis might be limited by the inability of the plants to uptake enough nitrogen and phosphorus which are often limited in wetland habitats, quoting Mitsch et al. (1979), Day (1982) and Bowden (1987). However, Hetherington et al. (2016) have more recently shown that the Stigmaria roots are branched, developed at a density of c. 25 600 per metre and were covered in root hairs; Boyce & DiMichele (2016), not having access to this later publication, thought they had neither. Hetherington et al. (2016) suggested that these highly branched rootlets would have contributed to the anchorage of the arborescent lycophytes, but they would also have increased the absorption power of the Stigmaria roots and augmented the ability to absorb the necessary nitrogen and phosphorus for growth. There is one further factor which needs to be taken into account. All of the well-preserved Stigmaria described by Thomas & Seyfullah (2015b) were ‘rooted’ in sediments and not in coal where they would have access to minerals. Clearly not all Stigmaria grew in sediments because fragments have been found encased in coal balls (e.g. Frankenberg & Eggert, 1969) and the bases of trunks are known to occur in shales immediately above coal seams. Such fossilized trunks known as kettle bottoms can be hazardous structures in coal mines because they are liable to fall out of the mine roof without warning. There is a possibility that the differing positions of growth of the Stigmaria reflect dissimilar arborescent lycophytes adapted to distinctive ecological conditions. It is clearly impossible to be precise about the length of time it took arborescent lycophytes to grow and reach fertile maturity. Indeed, it is highly probable that some plants became fertile long before others. The large ‘ulodendroid’ scars that are borne by some stems such as Lepidodendron Sternberg (e.g. Lepidodendron (e.g. 1966, (e.g. and Ulodendron majus Lindley and Hutton (e.g. Lindley & most show the positions of branches that may have borne Such a growth pattern would have cones while the plant was still with a crown of dichotomizing branches would have size before fertile by cones on their terminal after the plants size and producing cones they must have and started to otherwise they would have the development of the generation of plants through and the of photosynthesis by In the of the arborescent lycophytes is most with them having been very fast growing trees, as suggested by Phillips & DiMichele This is also in with the economic construction of the stems with little secondary the large number of stomata that they the of the and the relatively well-illuminated in which the young plants We cannot with the recently by Boyce & DiMichele (2016) that they grew much more some of the key in their and of stomata and the structure of the stigmarian root have since been challenged by Because of their with living plants to be In the plants would have been to reach in much less time than conifer and angiosperm trees, and less than the of years suggested by Boyce & DiMichele