The invasion of the land by plants: when and where?

Abstract

The origin of land plants was one of the most important events in the history of life on Earth. It was a major macroevolutionary event in its own right, with profound ecological consequences, but it also had enormous effects on the environment of planet Earth, altering atmospheric composition, weathering and soil formation, etc., and hence climate and biogeochemical cycles. Understanding the timing of the origin of land plants is a long term goal. In this issue of New Phytologist, Rubinstein et al. (pp. 365–369) provide new evidence that this event occurred 8–12 million yr earlier than previously accepted. ‘Thus, although reports are currently few, attention is turning to the possibility that the centre of origin of land plants may have been located on Gondwana.’ The land plants (Embryophytes) are a monophyletic group that evolved as an adaptive response to the migration from a freshwater aquatic to terrestrial subaerial habitat. Phylogenetic analysis of extant plants suggests that charophycean green algae share a sister group relationship with the Embryophytes, that is, the land plants probably evolved from a freshwater aquatic multicellular green alga similar to extant Chara and Coleochaete (Graham, 1993). Within the Embryophytes liverworts are the most basal group, followed by mosses, and then hornworts and vascular plants sharing a sister group relationship (Qiu et al., 2006). However, it is to the fossil record we must turn if we are to understand what the first land plants were like and when and where they evolved. Traditionally the earliest evidence for land plants was actual megafossils (fossils representing a significant portion of the plant). Until the late 1950s the simple rhyniophytoid plant Cooksonia provided this benchmark (Lang, 1937), and it is still the oldest generally accepted megafossil, being reported from the Late Silurian (late Wenlock) (Edwards et al., 1983). However, evidence from a new technique called palynology became widely available from the late 1950s. This technique involves dissolving rock to release small organic particles trapped within it as it was deposited. Among these particles are the subaerially released spores of land plants. All land plants produce spores, or their homologue pollen, in vast numbers. They are enclosed within a thick resistant wall that protects them from physical abrasion and exposure to UV-B radiation during dispersal. Upon release, spores/pollen are widely dispersed by wind (or other vectors such as insects) before eventually being deposited, often after remobilization by water. Thus, spores/pollen tend to accumulate wherever sediment is being deposited. Sporopollenin is one of the most resistant organic macromolecules known to man and readily fossilizes. Thus palynology yields a remarkably complete fossil record of the dispersed spores/pollen of land plants. However, it is important to note that spores/pollen enclosed in a resistant sporopollenin wall are a synapomorphy of land plants (Wellman, 2003). When plants migrated out of water and into the subaerial environment sexual reproduction became problematic as reproductive structures could no longer simply be released into the aquatic environment to swim to one another. Rather, they must be transported through the harsh subaerial environment. Thus sporopollenin-coated spores/pollen are absolutely essential for subaerial existence as a land plant, but of little or no use to an aquatic plant (in fact, secondarily aquatic ‘land plants’ generally lose the sporopollenin coat of their spores/pollen). An early palynological discovery was that the fossil record of trilete spores (similar to the isospores of homosporous plants such as many modern lycopsids and ferns) extends back into the earliest Silurian (Hoffmeister, 1959). This predated the earliest land plant megafossils and provided a new benchmark for the origin of land plants (Fig. 1). It also suggested that the land plant megafossil record was patchy (i.e. incomplete and biased) as might be expected: megafossils are produced in smaller numbers and are much less likely to survive the fossilization process than dispersed spores/pollen. Geological time scale for the Ordovician–Silurian illustrating benchmark palaeontological studies relevant to dating the origin of land plants. Figures refer to ages in millions of years. A further palynological shockwave was provided by Gray & Boucot (1971). They described spores that occur in unusual configurations (now termed cryptospores) in even older rocks, subsequently shown to extend down into the Ordovician (Gray, 1985). This work was at first highly controversial. Many claimed that these spores did not represent land plants. After all, where were the plant megafossils? However, supporting evidence has subsequently accumulated to the extent that land plant affinities are almost universally accepted. It was also realized that the earliest land plants were probably ‘bryophyte-like’ (liverworts are the most basal extant land plants). Bryophytes lack abundant recalcitrant (e.g. lignified) tissues and consequently do not fossilize easily and have a negligible fossil record. This probably explains why the earliest land plants left a rich dispersed spore record but only a very limited megafossil record (Wellman et al., 2003). The evidence for land plant affinities of cryptospores may be summarized as follows: Cryptospores are morphologically similar to land plant spores in terms of size and possession of a thick, resistant wall (see (7) below). However, they occur as monads, dyads and tetrads, rather than strictly as monads formed from the dissociation of a meiotic tetrad. Furthermore, they are often enclosed within a thin envelope that is difficult to equate with similar structures in extant plants (e.g. Richardson, 1996) (Fig. 2). Cryptospores occur in terrestrial deposits. They are also recovered from marine deposits where they decline rapidly in abundance offshore (as do modern spores/pollen flushed into the sea). Phylogenetic analyses in the mid-1980s suggested that liverworts are the most basal extant land plants. Certain cryptospores (notably permanent tetrads enclosed within an envelope) resemble the spores of certain extant liverworts (Gray, 1985). Taylor (1995) demonstrated that some cryptospore dyads have spore wall ultrastructure only known among extant liverworts. Remarkably preserved Lower Devonian plants contain in situ cryptospores and these plants have certain bryophytic characters (Edwards et al., 1999). Complete sporangia (including covering) with in situ cryptospores were recovered from the Ordovician (Katian) of Oman (Wellman et al., 2003). Those containing dyads also have spore wall ultrastructure known only among extant liverworts. Recent geochemical analysis has demonstrated that the spore wall in cryptospores is chemically similar to that of known land plant spores (Steemans et al., 2010). Cryptospores from the Late Ordovician (Katian) of Oman. (a) Naked dyad; (b) dyad enclosed in a thin walled envelope, part of which is discernible on the edges of the specimen; (c) naked tetrad; (d) tetrad enclosed in a thin-walled envelope, part of which is discernible on the edges of the specimen. All specimens approx. 30 microns in maximum diameter. Until now the earliest reported cryptospores were from the Darriwilian (Middle Ordovician) from the Czech Republic and Saudi Arabia (Strother et al., 1996). Interestingly both of these reports concern the large Ordovician supercontinent of Gondwana and its margins. Now Rubinstein et al. provide an even earlier report (by some 8–12 million yr) of cryptospores from the early Middle Ordovician (Fig. 1). This provides a new benchmark for the origin of land plants. Their report is from Argentina which was also situated on the supercontinent of Gondwana. Thus, although reports are currently few, attention is turning to the possibility that the centre of origin of land plants may have been located on Gondwana. This mirrors the situation whereby the centre of origin of vascular plants is currently hypothesized to have occurred on this continent (Steemans et al., 2009). Understanding the origin of land plants is a multidisciplinary effort with diverse strands of evidence derived from palaeontological, biological and geological investigation. Palaeontological research continues to throw up new finds, like that reported by Rubinstein et al., that modify an emerging picture of the nature of the earliest land plants, where and when they first appeared, and how they colonized the Earth’s surface. However, much work remains to be done as we struggle to unravel the vagaries of the fossil record and many controversies persist. There are claims that thick-walled palynomorphs from the Cambrian represent an even earlier record for land plants. However, many interpret these as the resting cysts, or even body cells, of aquatic algae (Wellman, 2003). The jury is out and more evidence is required. Even more intriguing is the unusual configuration of the earliest land plant spores (cryptospores). What does the envelope represent? Why do they often occur in dyads? It is clear that most extant and fossil land plants produce spores/pollen in tetrads by meiosis and usually the tetrads dissociate into four monads before dispersal. Presumably, dyads are normal meiotic products but developed via separation between two successive meiotic divisions. But why did dyad formation operate among land plants for the first 65 million yr of their existence before entirely disappearing? Possibly two lineages of embryophyte existed and the dyad-producers were simply outcompeted by the tetrad producers owing to some other factor unrelated to reproduction. These, and other problems will be addressed by new evidence from a variety of disciplines. This will include new fossil finds, interpreted, as always, in the light cast by our understanding of extant plants. I suspect that this will be forced onward by the current revolution in evo-devo studies, fuelled by the proliferation of molecular data, including whole-genome sequences from more basal land plants such as Physcomitrella and Selaginella.