The nature and evolutionary relationships of the earliest land plants


Palaeobotanists have long speculated on the nature and evolutionary relationships of the earliest land plants. What were they like in terms of anatomy and physiology? How are they related phylogenetically to the green algal ‘ancestors’ that came before and the extinct/extant land plant clades that came later? In this issue of New Phytologist Edwards et al. (pp. 50–78) review aspects of more than a quarter of a century's research on some of the most spectacular early land plant remains ever discovered: the Late Silurian–Early Devonian Welsh Borderland Lagerstätten of early land plants exceptionally preserved by charcoalification. These are not the oldest fossilized land plants, but they are the oldest exhibiting exceptional preservation. Rare elements of these diverse floras have much to inform us regarding the nature and evolutionary relationships of the earliest land plants.

‘Lagerstätte, or exceptionally preserved biotas, are the jewel-in-the-crown of palaeontology.’

Extant plants are useful in understanding the earliest land plants, but they have their limitations. Their prime importance is to provide us with a phylogenetic framework. The first cladistic phylogenetic analyses of living plants were based entirely on anatomical characters. By the mid-1980s such analyses were indicating that: (1) land plants (embryophytes) were monophyletic; (2) the embryophytes were sister group to the charophycean green algae; (3) within the embryophytes the bryophytes were paraphyletic with liverworts basal and either mosses or hornworts sister group to the vascular plants (Mishler & Churchill, 1985). By the 1990s molecular sequence data became available for use in cladistic phylogenetic analyses of land plants. These data generally supported the earlier findings, but seemed to confirm that it was the hornworts that were sister group to the vascular plants (Qiu et al., 2006). For the last 30 yr these phylogenetic analyses have provided a conceptual framework within which we have considered the earliest land plants. It led to speculation that because liverworts were the most basal extant land plants the earliest land plants may have been similar and were ‘bryophyte-like’ or ‘liverwort-like’ (Gray, 1985).

The importance of fossils is that they provide information on the nature of the organisms that actually lived at any given time, despite the much publicized vagaries of the fossil record and the fact that fossils provide incomplete anatomical information and usually no molecular data. Unfortunately the early land plant megafossil record is rather poor. Wellman et al. (2003) reported spore-filled sporangia from the Ordovician, but full reconstruction of these plants is far from complete. By the Silurian we have more completely preserved land plants but they are known from only 24 localities worldwide (reviewed in Wellman et al., 2013). None-the-less, when fossils are incorporated into cladistic phylogenetic analyses of land plants the pattern observed earlier tends to hold up (Kenrick & Crane, 1997). Intriguingly, however, the Silurian assemblages lack clear evidence for bryophytes but contain primitive vascular plants (rhyniophytes and zosterophylls). The earliest fossil bryophytes are from the Early Devonian, and these are strangely modern in aspect, and can be attributed to the Jungermanniopsida: Metzgeriales (Guo et al., 2012). This is not what one would expect based on the generally accepted phylogenies. Such a discrepancy is usually attributed to the poor preservation potential of bryophytes, although this is becoming more difficult to accept based on the discovery of well preserved bryophytes from the Devonian and the exceptional preservation of plants in Lagerstätten such as the Welsh Borderland charcoalified material and the Rhynie chert.

Lagerstätte, or exceptionally preserved biotas, are the jewel-in-the-crown of palaeontology. They form when organisms are preserved in unusual environments and conditions in an exceptional way, either in unusual numbers, articulation or preservation (e.g. labile tissues are preserved in addition to the normally preserved recalcitrant tissues). They are important because they are so much more informative regarding the anatomy of an organism. The vast majority of early land plant fossils are preserved as carbonized compressions. These form when a plant is trapped in sediment and buried. As pressure and heat increases over time, the plant is fossilized as it is compressed and transformed to coal as volatiles are driven off. Such fossils provide details of morphology but only limited anatomical information. For example, early land plants may preserve stomata, conducting tissues, in situ spores, etc., but little cellular detail remains and the majority of labile tissues are lost. However, palaeobotanists are extremely fortunate that several early land plant Lagerstätten have been discovered. One is the famous Early Devonian (Pragian) Rhynie chert from Scotland where the plants are perfectly preserved by silicification uncompressed and with complete cellular detail intact (Kidston & Lang, 1917).

The Rhynie chert flora, although much younger than the earliest land plants (Fig. 1), has several noteworthy features pertinent to the discussion herein. First, there are no bryophytes represented in the flora. Second, the plants exhibit rather unfamiliar character combinations. For example, gametophytes are larger than anticipated and possess internal conducting tissues and stomata (Kerp et al., 2003). But, as discussed earlier, one of the strengths of the fossil record is that it reveals extinct characters and unfamiliar character combinations that existed in the past. Evidence for these is not available from the study of extant plants alone. To accommodate such fossils palaeontologists formulated the concept of crown, stem and total groups (Jefferies, 1979). This essentially involved recognition that the numerous characters that define crown groups were often acquired incrementally over a protracted period of geological time.

Figure 1.

Geological timescale for the Middle Ordovician–Early Devonian illustrating the ages of the Lagerstätten and important evolutionary events discussed in the text. Numbers refer to ages in millions of years.

At this juncture we must take a step backwards and consider the very earliest evidence for land plants. Early land plants have two fossil records: plant megafossils that preserve a significant portion of the plant and dispersed spores that are released in their hundreds of thousands by the plants during life. Dispersed spores have very high preservation potential and form a continuous fossil record of early land plants from the Mid Ordovician (reviewed in Wellman, 2010). They indicate that there was a cryptic phase of land plant evolution for some 35 million yr from the Mid Ordovician to Early Silurian before primitive vascular plants appeared in significant numbers in the Early Silurian. Gray (1985) suggested that the somewhat unfamiliar spores that occur during this period (cryptospores) were produced by the most primitive land plants that were ‘bryophyte-like’. But without a megafossil record of these plants we have little idea what they were like and where they sit with respect to the stem lineage leading to crown group plants. This is why the material described by Edwards et al. is so important.

The two main Lagerstätte described in this issue of New Phytologist by Edwards et al. are both from the Welsh Borderland and are Ludford Corner (Late Silurian: Pridoli) and north Brown Clee Hill (Early Devonian: Lochkovian). Both preserve early land plants semi-compressed and in astonishing cellular detail because they were charcoalified during wild fire burning (Edwards & Axe, 2004). The assemblages are highly diverse with numerous identified plant taxa. The majority consist of bifurcating axes with terminal sporangia containing in situ trilete spores (rhyniophytes with evidence for vascular tissue and rhyniophytoids with no definitive evidence for vascular tissue). However, among these abundant fossils some rare specimens contain in situ cryptospores, rather than trilete spores, and have been termed cryptophytes (Edwards et al., 1995). Although these fossil plants are younger than the main phase of cryptospore evolution (Fig. 1) it appears they may represent holdovers from this flora and provide a unique insight into what these plants were actually like. Edwards et al. carefully document the unfamiliar anatomical character combinations displayed by these rare plants and attempt to ascertain where they are placed among the stem group embryophytes in terms of phylogenetic relationships.

Despite huge leaps in progress in recent years we are still far from understanding the nature and evolutionary relationships of the earliest land plants. Chief among our problems is the lack of fossil evidence from the Ordovician. Other than the scanty remains described by Wellman et al. (2013) we having nothing but spores to show for the first 35 million yr of land plant evolution. The work of Edwards and colleagues on younger material offers some hope of understanding these plants, but we desperately need to discover an Ordovician Lagerstätte to plug this knowledge gap. Then there are concerns regarding our understanding of land plant phylogeny. We have a generally accepted phylogenetic hypothesis for extant land plants that has remained more-or-less intact for 30 yr. But just how robust is this phylogeny and how do the unfamiliar fossils of the earliest land plants fit into it? We desperately need a combined analysis that incorporates evidence from both living and cryptophyte fossil plants, because the unusual character combinations present in early land plants from the Welsh Borderland and Rhynie chert Lagerstätten hint that there may be another way. Could the bryophytes be monophyletic, with both the bryophytes and vascular plants having evolved from plants with prominent gametophytes and sporophytes, as preserved in the Rhynie chert? Such a hypothesis is beginning to be explored seriously, with evidence coming from both fossils and extant plant molecular data (Goremykin & Hellwig, 2005).