The generation of new seafloor at mid-ocean ridges is counterbalanced by the subduction of old seafloor beneath continental plates. Water released from these plates as they sink into the mantle lowers the melting point of the rock above them and causes magma to rise to the surface, erupting to form volcanic mountains such as the Rockies and Andes. That is how new land forms now, and how it accreted around the ancient cratons in the Palaeozoic. It took time for plant and animal communities to establish themselves on this barren, initially steaming hot land. Here we explore some of the evidence for the first stages in the process. Which better accounts for the fossil sequence: recolonisation or step-by-step evolution?
Initially, the process may not have been dissimilar to the colonisation of volcanic islands that form today. Since the formation of new land on the planet has never totally stopped, there are opportunities to observe the sequence of events from the moment the lava solidifies. The colonisation of existing land after it has been resurfaced by volcanic eruptions also offers insights. Though modern eruptions are puny compared to the cataclysm that caused the entire planet to be resurfaced, it is evident that vegetation greens up the land in stages, not all at once.
Bacteria, fungi and algae are the first to build communities. They release chemicals which break the rock into grains. Carbon dioxide dissolved in rainwater creates a weak acid which reacts with the rock minerals, washing some away in solution and turning the residue into clay. Soil bacteria fix nitrogen into compounds which are useful to plants. Roots and humic acids break up the ground further and enhance the weathering process. In due time, burrowing ants and earthworms join the ferment. Over months and years, progressively greater depths of rock are converted into fertile soils. The succession of events is ecological, not evolutionary. Micro-organisms do not evolve into mosses, or mosses into ferns. Each group has an independent origin from beyond the area colonised.
The correspondence between modern ecological successions and the order of fossils is not perfect. In important respects the yet-to-be-vegetated world was different from the modern world. Rivers in the modern sense of meandering long-lived low-energy channels cutting into a vegetated landscape did not exist. Alluvial landscapes in the Cambrian consisted of broad, unconfined blankets of coarse sediment, not organic-rich muds (Davies & Gibling 2011). Earthquakes and gravity slides were continually eroding the mountains and producing ever thickening sheets of sand upstream and silt downstream. Flash floods were generated not so much by rain-storms as by pulses of water surging through faults, as they suddenly slipped and brought up water from beneath the crust. Plants could not establish themselves so long as surfaces were continually being buried.
As conditions became quieter through the early Palaeozoic, the proportion of mudrock to sandstone increased, while the overall size of the sand grains decreased (Davies & Gibling 2010). Pauses between one depositional event and another lengthened, until they were just long enough for simple plants to become established: plants that grew quickly and did not require a large investment of biological energy. Thus, for some while, plant fossils tended to represent only pioneer species, at the beginning of potential successions. Although diversity was increasing, species that took successions beyond the pioneer stage only stood a chance of impressing the record when a locality had been stable long enough for them to take root and was then overtaken by a depositional event rapid enough for the vegetation to be fossilised.
Nor were new stocks of vegetation ‘just round the corner’. In the case of Surtsey, a volcanic island that formed off Iceland in the 1960s, plant seeds were brought over in the gizzards of migrating birds. Birds were scarce in the Palaeozoic – they are rarely fossilised anyway, but none from the Palaeozoic have been preserved as fossils – and many of the plant species characteristic of modern ecosystems had not yet evolved. On an earth denuded of fauna and flora, there was no possibility of bringing in relief supplies from somewhere nearby. Ecological systems had to recover from scratch, in an uphill struggle where attempts at recovery were repeatedly being disrupted.
The places where vegetation would most easily have become established were those furthest away from the highlands, such as low-lying plains and deltas. Here the energy of rainstorms and discharges of subterranean water was mostly dissipated and the rate of deposition of silts and clays on the margins of ephemeral river channels slow enough to give vegetation an opportunity. Tectonic processes in the mountains, coupled with the weathering effects of acid rain (at a time when there was much more carbon dioxide in the atmosphere), had already done the hard work of converting hard rock into grains. The successful plants were those suited to wetland environments, comparatively simple in design. They were not pioneers in the sense of preparing the ground for others to follow. They became established because they were semi-aquatic plants that did not require mature soils, ‘r-strategists’ that reproduced quickly and proliferated over wide areas.
Bacteria, fungi and algae reproduce by means of spores. So do certain plants, such as mosses, lycopsids (clubmosses), horsetails and ferns. Plant spores have very resistant walls and are easily transported by wind and water because of their lightness and smallness; their tough walls protect them against damage during transport, desiccation and attack by microbes. Consequently they are not only the best preserved plant remains in old sedimentary rocks but also the first, and by far the most abundant. By contrast, the spores of bacteria, fungi and algae are rarely preserved.
Spores of mosses and liverworts are known from at least the mid Ordovician onwards, less certainly from the Cambrian, and their abundance increases with ascending stratigraphic level, as one would expect if plant fossils were reflecting the progressive recovery of vegetation. Macrofossils before the Devonian are also scarce because conditions were unfavourable for their preservation. In particular, the Ordovician-early Silurian was a time of persistently high sea-levels, leaving few continental deposits. Many of the earliest spores occur in nearshore marine deposits, blown out to sea by the wind.
Though simpler than most plant types, mosses and liverworts are intrinsically complex, as is apparent from their life cycle and from their ability to photosynthesise. As with all plants and some algae, their generations alternate between haploid phases, plants with one set of chromosomes, and diploid phases, plants with two. How they fit into the evolutionary tree of life is unknown. They each constitute separate lineages and are neither preceded by evolutionary forerunners nor followed by evolutionary successors. Capable of withstanding withstanding huge fluctuations in moisture and temperature, mosses and liverworts can survive in habitats inimical to other plants. They have changed little over time because they are so well-designed for their particular ecological role. Nonetheless, although unchanging as a group, they have become extremely diverse. Almost 19,000 species are known today.
Mosses were among the first plants to grow on Surtsey. Within a few years they were growing around crevices and holes where steam kept the volcanic rock damp. Nitrogen-binding cyanobacteria were also found around the steam-holes. Just seven years after the first eruption in 1963, mosses and lichens were widespread on the bare lava.
Fungi are not plants but multicellular organisms in a kingdom of their own. Their origin is as much an evolutionary mystery as that of the other kingdoms. More than 100,000 species are known, some aquatic, most of them terrestrial. They play a vital role as decomposers, recycling organic remains back to the environment in forms digestible by other organisms. Nearly all plants depend on symbiotic fungi called mycorrhizae to help the roots absorb minerals and water from the soil, while the fungi benefit by receiving carbohydrates from the plants.
Until recently the oldest traces of terrestrial fungi were filaments and spores dated to the Ordovician, about the same time as plant spores appear. They bear a strong resemblance to modern mycorrhizae that form filaments in or among roots in order to increase the surface area for absorption. Fungal-looking filaments also occur in the Silurian, with some again looking like extant species. Numerous types of fungi have been observed among the plants in Scotland’s Rhynie Chert, of early Devonian age. Shales in the DRC interpreted as collecting in ephemeral ponds preserve the remains of fungi 300 million years older than the Ordovician examples. Such records are always prone to be broken, because the ages reflect accidents of discovery and preservation, not the time at which a group can be said to have evolved.
Lichens consist of fungi living in partnership with cyanobacteria or algae. DNA analysis of extant forms has led to the conclusion that lichens arose independently five times or more! The fossil record of lichens is poor, but cyanobacteria and algae first appeared in the Archaean, fungi probably in the Proterozoic, and there are reports of soils with lichen-like fossils dating back to the Cambrian. The oldest unequivocal lichen is Spongiophyton minutissimum, a widespread early Devonian fossil.
Plant photosynthesis involves the absorption of carbon dioxide and the emission of oxygen, mediated through pores. Since large amounts of water vapour are lost in the process, plants higher than a few centimetres require specialised tissues to draw water from the ground. In vascular plants these tissues are of two types: xylem, which pipes up water and minerals from the roots to the leaves, and phloem, which distributes sugar and other products of photosynthesis from the leaves to the roots. Xylem also provides stems with rigidity, another requirement if plants are to grow higher than a few centimetres. The appearance of these tissues was one of the most significant events in the history of land plants. Some mosses also have conducting cells, but these originated independently from those in vascular plants.
How natural selection acting on genetic mutations leads to new designs is far from clear, and assuming it did, one would expect it to give rise to no more than adequate design. The characteristics of distinctly good design, such as energy efficiency or ingenuity in the solution of a given problem, are evidence of an intelligent creative power. In this instance the design is superb. ‘It is difficult to imagine a cheaper process for driving the transpiration stream’, writes John Sperry. The coupling between evaporation at the plant surface and the negative pressure achieved by capillary forces in the cell walls produces the driving force almost free of charge. Xylem conductivity per area exceeds that of non-vascular plants by six orders of magnitude.
The minimum requirement for such a system is a genetic program that causes the death of the cells lining the conduits and the manufacture and deployment of a substance (lignin) to strengthen the cell walls against collapse. These innovations were already in place by the late Ordovician, when vascular plants left evidence of themselves in the form of characteristic spores. The oldest known whole vascular plant, Cooksonia, did not appear until the mid Silurian, when the processes of erosion and deposition around them were presumably now less destructive. Cooksonia was small (less than 10 cm high), simple in appearance, and like the spores initially rare. By the end of the period (supposedly millions of years later) it was globally widespread, though still rare, and comprised at least five species.
The relationship of Cooksonia to the later vascular plants is as little understood as the origin of the whole group. This later vegetation included zosterophytes (from the Greek for garland, describing the successively arranged sporangia, and phyton, meaning plant), rhyniophytes (named after the Scottish village of Rhynie), trimerophytes (referring to their multiple branching, as illustrated by Psilophyton below) and lycopods (referring to the resemblance of some branch tips to a wolf’s paw). Most did not have leaves and roots, but the lycopod Baragwanathia had leaves (as do mosses), and all of them branched upwards off horizontally growing stems called rhizomes, which anchored the plants. In other respects they differed markedly from each other, notably in the design of their conducting walls, which were more complex and various even than those of extant plants. Since these diverse groups all appeared about the same time (late Silurian-early Devonian), a shared ancestry seems unlikely. The zosterophytes, rhyniophytes and trimerophytes died out in the course of the Devonian and left no descendants. A few lycopod species exist today, though much diminished both in size and diversity.
Baragwanathia, the earliest examples of which come from Australia, is especially problematic because it is taller and more ‘advanced’ than its fossilised contemporaries. Other anomalously early appearances include vascular plant fossils from Late Silurian deposits on Bathurst Island, Canada (Kotyk et al. 2002), and Early Devonian deposits around Gaspé Bay, Canada. In the latter case, the plants achieved a stature of 2–3 metres and were fossilised in situ, showing that they were capable of rooting to a depth of nearly 1 metre (Elick et al. 1998). Despite similarities in anatomy (the presence of a root cap, root hairs and an endodermis), there is no evidence that the roots of lycopods and non-lycopods evolved from a common ancestor (Hetherington & Dolan 2018).
Because of these and many other discoveries the ‘somewhat simplistic’ picture of plant evolution has had to change (Edwards & Richardson 2004). Simplicity of outward form masks complexity at the cellular level. Almost from the moment they appear, vascular plants are diverse, making it difficult to see how, invisible to the fossil record, they might all have arisen from a single vascular plant ancestor. The gap between vascular and non-vascular plants such as the mosses and liverworts is even larger. So is the gap between all land plants and their closest presumed relatives, the green algae (Kenrick 2000). By contrast, the time gap between the vascular and non-vascular plants is comparatively small. Mega-evolution simply is not an appropriate inference to draw.
While it is recognized that the assemblages provide the most complete and extensive record of the history of vascular plants in a restricted geographical area during the time interval, it seems likely that major evolutionary innovation occurred elsewhere.
‘Major evolutionary innovation’ always occurs off camera. What we actually see is organisms that have already been ‘innovated’ – one might say, created – parachuting in from somewhere else, colonising virgin territory. Currently, vascular plants appear to have originated in Gondwana, i.e. present India and the southern continents (Anderson et al. 1999, Wellman et al. 2013).