Journal of Experimental Botany

This Article Abstract FREE Full Text (PDF) E-letters: Submit a response Alert me when this article is cited Alert me when E-letters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Disclaimer Google Scholar Articles by Raven, J. A. Articles by Edwards, D. Search for Related Content PubMed PubMed Citation Articles by Raven, J. A. Articles by Edwards, D. Agricola Articles by Raven, J. A. Articles by Edwards, D. Social Bookmarking          What's this?

Journal of Experimental Botany, Vol. 52, No. 90001, pp. 381-401, March 2001
© 2001 Oxford University Press

Roots: evolutionary origins and biogeochemical significance

J. A. Raven^1,3 and D. Edwards²

¹ Division of Environmental and Applied Biology, School of Biological Sciences, University of Dundee, Dundee DD1 4HN, UK
² Department of Earth Sciences, Cardiff University, PO Box 914, Cardiff CF1 3YE, UK

Received 1 June 2000; Accepted 20 September 2000

	Abstract

Top Abstract Introduction What is a root? The functions of roots... The biogeochemical context of... The early (Palaeozoic) evolution... Implications for plants and... Conclusions References

 
Roots, as organs distinguishable developmentally and anatomically from shoots (other than by occurrence of stomata and sporangia on above-ground organs), evolved in the sporophytes of at least two distinct lineages of early vascular plants during their initial major radiation on land in Early Devonian times (c. 410–395 million years ago). This was some 15 million years after the appearance of tracheophytes and c. 50 million years after the earliest embryophytes of presumed bryophyte affinity. Both groups are known initially only from spores, but from comparative anatomy of extant bryophytes and later Lower Devonian fossils it is assumed that, during these times, below-ground structures (if any) other than true roots fulfilled the functions of anchorage and of water and nutrient acquisition, despite lacking an endodermis (as do the roots of extant Lycopodium spp.). By 375 million years ago root-like structures penetrated almost a metre into the substratum, greatly increasing the volume of mineral matter subject to weathering by the higher than atmospheric CO₂ levels generated by plant and microbial respiration in material with restricted diffusive contact with the atmosphere. Chemical weathering consumes CO₂ in converting silicates into bicarbonate and Si(OH)₄. The CO₂ consumed in weathering ultimately came from atmospheric CO₂ via photosynthesis and respiration; this use of CO₂ probably accounts for most of the postulated 10-fold decrease in atmospheric CO₂ from 400–350 million years ago, with significant effects on shoot evolution. Subsequent evolution of roots has yielded much-branched axes down to 40 µm diameter, a lower limit set by long-distance transport constraints. Finer structures involved in the uptake of nutrients of low diffusivity in soil evolved at least 400 million years ago as arbuscular mycorrhizas or as evaginations of ‘roots’ (‘root hairs’).

Key words: Devonian, endodermis, phylogeny, Silurian, weathering.

	Introduction

Top Abstract Introduction What is a root? The functions of roots... The biogeochemical context of... The early (Palaeozoic) evolution... Implications for plants and... Conclusions References

 
The evolution of ‘the root’ has attracted much less attention than that of ‘the shoot’. This is particularly true of the evolutionary origin of the root, where the limited economic significance of the lower vascular plants is one of the factors constraining consideration of this topic. Furthermore, roots may be less readily fossilized than shoots, and the fossil record is, as we shall see, not very helpful. The position was well put in the following quotation from Gifford and Foster. ‘The differentiation between distinct root and shoot systems in most living vascular plants is of considerable interest from an evolutionary viewpoint. It is now rather generally agreed that this differentiation did not exist in such Devonian land plants as Horneophyton and Rhynia. In these archaic organisms portions of the underground system of stems apparently served physiologically as roots. It seems reasonable to postulate that roots were acquired later in the evolution of vascular plants. However, the steps in the evolutionary divergence of roots from primitive shoots, which led to the acquisition of a root cap, a prevalent internal or endogenous origin of roots, and the retention of a primitive type of vascular system, are unfortunately obscure today’ (Gifford and Foster, 1987). More recently, Fitter and Moyersoen wrote, ‘Complex root systems have been a feature of land plants for a long time, although the fossil record is unclear as to exactly how long’ (Fitter and Moyersoen, 1996).

This review considers root evolution by attempting to define the root and, in so doing, delimiting its difference from the shoot, and by considering the functions of the roots of extant plants and cases in which ‘root’ functions are performed by non-root structures. The article then considers the fossil history of vascular plants and the environment in which they evolved, the time at which true roots first appear in the fossil record, and the role of roots and root-like structures which fulfil root functions in modifying global biogeochemistry and its influence on shoot evolution.

	What is a root?

Top Abstract Introduction What is a root? The functions of roots... The biogeochemical context of... The early (Palaeozoic) evolution... Implications for plants and... Conclusions References

 
The anatomical and morphological material assembled here comes from many sources (Bell, 1991; Bierhorst, 1971; Cutter, 1971; Eames, 1936; Eames and MacDaniels, 1947; Esau, 1953; Fahn, 1974; Gifford and Foster, 1987; Goebel, 1930; Groff and Kaplan, 1988; Raven, 1977, 1984a, b, 1986, 1993, 1995a, 1997).

Roots are axial multicellular structures of the sporophytes of vascular plants which usually occur underground, have strictly (root cap notwithstanding) apical elongation growth, and generally have gravitropic responses which range from positive gravitropism to diagravitropism, combined with negative phototropism. Developmentally, the apical meristem of one (most lower vascular plants) to many (all seed plants) dividing cells produces a root cap acropetally and initials of the stele, cortex and epidermis basipetally. The branching of roots involves the endogenous origin of new root apical meristems in the pericycle or, in many ferns (e.g. Marsilea), the endodermis of the parent root. The lateral roots emerge from the pericycle through the endodermis or from the endodermis (when present, as it usually is), then pass through the cortex and the epidermis. Roots do not have superficial production of cells or tissues other than unicellular enations (root hairs). The primary stele of roots is generally a protostele (i.e. a solid cylinder of xylem, with or without parenchyma cells), often with convolutions of the stele–pericycle interface to produce an actinostele. There are roots which have a central pith with peripheral vascular tissue within the endodermis; examples are most monocotyledonous angiosperms. Roots do not have the segmentation of the axis between (superficial) leaves on the stem which is typical of shoots (stems plus leaves plus sporangia and their modifications such as strobili and flowers and their derivatives, seeds and fruits) to produce internodes between nodes. Roots arise from the embryonic axis, from other roots or stems; the latter is where all roots originate in non-seed vascular plants.

The root cap is thought to be a source of lubrication to the growing root apex, and a means of causing adherence of mature roots to soil particles via the production of mucilage, while living detached root cap cells (root border cells) may act as a means of manipulating the microbiota of the rhizosphere (McCully, 1999; Hawes et al., 2000). The root cap is also involved in graviperception.

Shoots, which are the other organ system of vascular plants, contrast with roots by being typically above-ground and their gravitropism ranges from negatively gravitropic to diagravitropic. The shoot apical meristem does not produce a (root) cap. The branching of shoots, including the production of leaves, is superficial (exogenous). The primary vascular tissue is sometimes a protostele, more usually a siphonostele or dictyostele (i.e. both with a pith, and a more or less dissected hollow cylinder of vascular tissue). An endodermis is less frequently present in shoots than in roots. Shoots growing in air can (as can the much less common aerial roots) have intercalary as well as apical elongation growth since they do not have the buckling problems which below-ground organs with intercalary growth would have as they grow through a much more resistant medium than air (Gill and Tomlinson, 1975). Shoots with intercalary growth growing negatively gavitropically, or diagravitropically can have problems with posture maintenance since by definition an extension zone must rely on a hydrostatic rather than a rigid skeletal system; aerial roots which grow downwards have no such problems. Shoots can generally give rise to roots (as well as shoots); roots more rarely give rise to shoots (Groff and Kaplan, 1988).

The distinguishing features of the roots (as compared to the shoots) of vascular plants are thus the occurrence of a root cap, a more defined lineage of cells from the apical cell(s) to tissues in the more mature parts of the roots, the essentially universal occurrence of an endodermis, a protostele (sometimes with a pith) and endogenous origin of lateral roots from roots (Table 1). Many root systems show localized root proliferation in response to spatial variability in resource availability (Robinson, 1996; Robinson et al., 1991, 1998; Zhang and Forde, 1998).

View this table: [in this window] [in a new window]   Table 1. Relationships of characteristics among the ‘original axis’ (above- and below-ground) of early vascular plants, and the shoots and roots of extant plants Data from Bell (1991), Cutter (1971), Damus et al. (1997), Duckett et al. (1996), Eames (1936), Eames and MacDaniels (1947), Esau (1953), Fahn (1974), Gifford and Foster (1987), Goebel (1930), Groff and Kaplan (1988), Raven (1977, 1984a, 1986, 1993, 1995a, 1997), and Schneider (1996).

 
Groff and Kaplan, in their excellent critical review of the concept of roots versus shoots in vascular plants (Groff and Kaplan, 1988; see also Goebel, 1930), very wisely discourage the use of the term ‘adventitious’ as applied to roots, since it can imply that the production of roots by stems is always abnormal, whereas it is the norm in many plants. Similarly, the use of ‘primary’ is discouraged in the context of root systems, since it can mean either ‘the first root’ or ‘the persistent root’ (Groff and Kaplan, 1988). These authors also counsel that the production of shoots from roots is not to be regarded as anomalous.

In the context of production of shoots by roots, it is instructive to consider N₂-fixing leguminous root nodules; these are essentially shoot-like structures which are produced by the roots of many members of the Fabaceae in the presence of appropriate rhizobia (Sprent, 1989; Mathesius et al., 1998; Gualtieri and Bisseling, 2000). The legume nodules develop from primordia that are initiated in the cortex and have peripheral vascular bundles and central symbiotic tissue. However, it is important to note that some legumes can form rhizobial nodules on shoot (stem) tissue as well as on roots (Parsons et al., 1993, 1995). Furthermore, the same rhizobial signal molecules (Nod factors) can induce the production of N₂-fixing nodules on the roots of the non-leguminous angiosperm Parasponia (Ulmaceae); these nodules are root-like with central vascular tissue and peripheral (cortical) symbiotic tissue, although they have no root cap tissue (Gualtieri and Bisseling, 2000). These rhizobial nodules of Parasponia have developmental and anatomical similarities to the actinorhizal (Frankia-containing) N₂-fixing nodules of members of the Rosid 1 clade families Rosaceae, Eleagnaceae, Rhamnaceae, Betulaceae, Casuarinaceae, Myricaceae, Coriariaceae, and Datiscaceae (noting that the rhizobially nodulating families Fabaceae and Ulmaceae are also in the Rosid 1 clade): (Gualtieri and Bisseling, 2000). Thus, actinorhizal nodules have central vascular and peripheral symbiotic tissue, and originate from root pericycle tissue, but lack a root cap (Gualtieri and Bisseling, 2000).

These examples of the production of shoots by roots mean that the phylogeny and the ontogeny of roots do not abolish their potential to produce shoot tissue.

Finally, Groff and Kaplan (Groff and Kaplan, 1988) recall the concepts of allorhizy and homorhizy (Goebel, 1930). Allorhizy denotes the development of shoot and root at opposite ends of the embryonic axis, as in extant seed plants, while homorhizy denotes root development which is lateral with respect to the longitudinal axis of the embryo, as in extant lower vascular plants. Later this study returns to Goebel's concepts in the context of the monophyletic or polyphyletic origin of roots (Goebel, 1930).

	The functions of roots in extant plants in relation to the capabilities of other below-ground plant structures

Top Abstract Introduction What is a root? The functions of roots... The biogeochemical context of... The early (Palaeozoic) evolution... Implications for plants and... Conclusions References

 
What do roots do in extant plants?
In addressing this question it is necessary to examine the occurrence of roots in extant plants, and of how the functions of roots are fulfilled in rootless plants in habitats where root functions are needed. The traditionally recognized root functions are anchorage (Niklas, 1992; Ennos, 1993; Ennos and Fitter, 1992; Berntson, 1994), and the uptake and longitudinal movement of water and mineral nutrients (Clarkson, 1985; Varney and Canny, 1993; Marschner, 1995). Even when roots are well developed some of these root functions may be partly performed by other structures. Thus, in rhizomatous plants, the below-ground rhizome may have had a role in anchoring the plant, while the fungal components of mycorrhizas may supplement or replace root hairs as the root components which increase the area of absorptive surface per unit of below-ground biomass (see below) (Clarkson, 1985; Caldwell and Richards, 1986; Fitter, 1987, 1996; Moteetee et al., 1996; Smith and Read, 1997; Blackwell, 2000; Redecker et al., 2000).

There are clearly constraints on the optimization of the various functions of roots in relation to the resource costs of producing and maintaining the roots as a function of their structure and, especially, their diameter. Thus, the absorptive (and anchorage) functions are maximized per unit of root biomass by small root diameter (Silberbush and Barber, 1983; Boot, 1989; Clarkson, 1985). However, the analysis of Silberbush and Barber showed that an even more important root attribute in maximizing the uptake of slowly-diffusing solutes is root elongation rate (Silberbush and Barber, 1983). Thus, for a nutrient such as phosphate in well-aerated soil, with phosphate binding to oxidized iron minerals, and a very low diffusion coefficient for the solute in soil, root elongation rate is very important in determining uptake of phosphate since it determines access to unexploited volumes of soil. However, Table 2 shows that, for the range of diameters of below-ground roots for which data are available, the larger diameter roots have the most rapid elongation rate. While there seems to be no mechanistic basis for this correlation, its existence shows that root diameter and root elongation rate are not independent variables, at least for the plants (all cultivated grasses) in Table 2. It is of interest that the correlation of extension rate and root diameter for cereals in Table 2 contrasts with hypotheses and data on slow-growing ‘magnolioid’ roots with a diameter up to 1.5 mm in the absence of secondary thickening and with a great dependence on mycorrhizas in the flowering plants (Baylis, 1975; St John, 1980). These are contrasted with faster-extending, smaller-diameter ‘graminoid’ roots with a greater potential to produce root hairs and less dependence on mycorrhizas (Baylis, 1975).

View this table: [in this window] [in a new window]   Table 2. Linear elongation rate of roots, root hairs, fungal hyphae and other structures as a function of the diameter of the organ and the medium (soil, air) in which elongations occurs Determinants of the elongation rate of roots are discussed by Barlow (1973), Canny (1973), Lyndon (1973), Passioura and Ashford (1974), Cahn et al. (1989), Varney et al. (1991), Bret-Harte and Silk (1994), and Pritchard (1994). The very rapid elongation of aerial roots of Cissus can be related to the large number of enlarging cells in the 1 m long elongation zone. This long elongation zone is permitted in aerial structures where buckling, such as would occur in below-ground structures with a long elongation zone, is not a consideration.

 
A further correlation with root diameter is that roots of smaller diameter may have higher construction costs and maintenance costs on a unit biomass basis than do larger diameter roots (Eissenstaadt and Yanai, 1997). This adds to the costs of producing and maintaining small diameter roots to set against the benefit, in nutrient-uptake terms, of a large surface area per unit biomass and the constraint on nutrient-uptake capacity due to their lower growth rate. Smaller diameter roots are also more likely to be determinate in their growth and to live for a shorter time than is the case for roots of larger diameter (Eissenstadt and Yanai, 1997; Skene, 1998, 2000). This also impacts on the mechanical effectiveness of roots, with the increased surface area for functional contact per unit biomass of smaller roots offset, in terms of mechanical benefit per unit construction cost, by their shorter lifespan.

After absorption of nutrient solutes and water by a root zone, the major fraction of the nutrients and water is transported axially toward the shoot in the xylem. Growth and maintenance of roots requires that photosynthate moves to them in the phloem. In both xylem and phloem mass flow of solution occurs along the conduits, and the Hagen-Poiseiulle equation shows that the flux of solution along a conduit under a given hydrodynamic driving force is proportional to the fourth power of the radius of the conduit (Fitter, 1987; Varney and Canny, 1993; Varney et al., 1991; Raven, 1999a, b). This constraint on axial flow in very small conduits imposes a lower limit on the diameter of roots, in that a root functional in axial conduction must have at least one xylem element and one phloem element, each with more than a certain minimum radius (Raven, 1999a, b).

Roots must also (with very few exceptions) have living cells in the stele associated with the xylem and phloem conducting elements, an endodermis, one or more layers of cortex, and an epidermis (Peterson, 1992; Damus et al., 1997; Aloni et al., 1998; Peterson and Cholewa, 1998; Raven, 1999a, b). In addition to these constraints on the minimum number of living cells in the radius of a root, there are also constraints on the minimum size of each of these living cells imposed by cost-benefit and other considerations (Raven, 1999a, b). Taken together, these constraints mean that the smallest radius of indeterminate roots of extant plants (20–25 µm; Fitter, 1987; McCully, 1999; Varney and McCully, 1991) can be rationalized, in functional terms, with roots of rather smaller radius possible if there is no endodermis (Damus et al., 1997).

While endodermis has been cited as an essential functional feature of roots, it is noted that the extant Lycopodium lacks an endodermis in its roots (but not in its shoots) (Damus et al., 1997), and that well-preserved fossil sporophytes of vascular plants (and other polysporangiophytes) do not have identifiable endodermis until the Carboniferous (Walton, 1940; Raven, 1984b, 1994b). Furthermore, such physiological work as has been performed on Arabidopsis thaliana mutants lacking a root endodermis (e.g. with mutation in the SHORT-ROOT (shr) gene) shows no significant functional impact on Ca transport to the shoot (Benfey et al., 1993; Scheres and Benfey, 1999; White et al., 2000).

This argument, as to the minimum size of multicellular underground absorptive and anchoring organs, has been couched in the context of the functional anatomy of roots. However, there seems to be no reason why stem (i.e. shoot components) structures could not be miniaturized to the same extent as the smallest extant roots, provided that they had a protostelic structure of their vascular tissue (Raven, 1999a). Clearly, an endodermis could be accommodated in such a small shoot radius just as is the case for roots. However, there seem to be no known below-ground shoot structures with as small a radius as these very small-radius roots. Any such small-radius stems with an endodermis occurring below ground could only with difficulty be distinguished from roots except for the absence of a root cap (Table 1), since these very small-radius structures are probably too small to branch, and thus be distinguished by exogenous rather than endogenous branching. Such small-radius stems would certainly have to be protostelic (Table 1).

Structure and function of root hairs and mycorrhizas
There are potential advantages in having diameters of nutrient-absorbing structures which are less than that consistent with the structure of roots. The argument here is an extension of that used above for roots. The uptake of a nutrient with low concentration and low diffusion coefficients in soil (e.g. phosphate in aerobic soils) has a higher uptake rate per unit biomass if that biomass is allocated to structures of smaller diameter (Silberbush and Barber, 1983; Clarkson, 1985; Boot, 1989; Harrison and van Buuven, 1995; Allaway and Ashford, 1996; Ashford et al., 1996; Eissenstadt and Yanai, 1997; Gilroy and Jones, 2000; Hodge et al., 2000; von Wirén et al., 2000).

Based on the orders of magnitude of the diameters of roots, root hairs and mycorrhizal hyphae in Read (Read, 1999; p. 171) a 2 µm diameter mycorrhizal hypha would have a 100 times greater surface area per unit biomass (volume), and a 20 µm diameter root hair 10 times greater surface area per unit biomass (volume), than a 200 µm diameter root. While these very small diameter ‘unicellular’ structures have advantages in terms of acquisition of resources such as phosphate by their increased surface area per unit biomass, they may have corresponding costs. One is that axial fluxes of solutes by cytoplasmic streaming may scale with cell diameter (Table 3). Another is that root hairs and mycorrhizas have a slow elongation rate (and thus a slow rate of exploitation of unexplored volumes of soil) (Table 2) and (at least for most root hairs) a short life-span, thus limiting the quantity of nutrient acquired per unit resource allocated to the structure (Eissenstadt and Yanai, 1997), at least as far as components (e.g. walls of root hairs) which cannot be recovered by the plant by disassembly and translocation to the longer lived parts of the plant are concerned (see below).

View this table: [in this window] [in a new window]   Table 3. Cytoplasmic streaming speed in relation to the diameter of the spherical or cylindrical cells in which it occurs; temperature 20–25 °C

 

Rootless rhizophytes: the function of roots performed by non-root structures in rootless organisms
Putting anchoring and absorbing structures into a wider context, it is important to remember that vascular plants rooted in soil on land belong to the large life form category known as rhizophytes (Raven, 1981, 1983, 1984a, b, 1986; Raven et al., 1998). These organisms, with plant structures in the soil or sediment which are large relative to the grain size of the soil or sediment, include a wide range of structures produced by photosynthetic organisms other than the roots (and root hairs and mycorrhizas) of terrestrial sporophytes of vascular plants. Even among the extant sporophytes of terrestrial vascular plants some (Psilotum, Tmesipteris) lack roots despite living in soil or epiphytically; here shoot structures with ‘root hairs’ and mycorrhizas are functional replacements for roots.

Roots are retained by most vascular plant sporophytes after they have moved back to live in water, even when they cease to be rhizophytes. Thus, while Utricularia and Ceratophyllum lack roots, the Lemnaceae usually retain roots despite being much-reduced pleustophytes (free-floating plants: Raven, 1981, 1999a), while the freshwater Podostemaceae and certain seagrasses (Phyllospadix spp.; some specimens of Amphibolis spp.) are haptophytes, with their roots modified as holdfasts with no special role in nutrient uptake, since haptophytes are benthic plants attached to substrata of a grain size much larger than the attaching organ, and often larger than the whole organism (Raven, 1981, 1999a). Among these extant vascular plant sporophytes the absence of roots is apparently a derived condition.

The free-living gametophytes of the homosporous free-sporing vascular plants are also rhizophytes, with the anchoring and absorptive structures comprising just the rhizoids (analogous to root hairs) of photosynthetic representatives, while non-photosynthetic taxa have fungal associates analogous to mycorrhizas (Raven, 1981, 1986, 1999a). No multicellular below-ground structures have been found in the gametophytes of now extinct Lower Devonian vascular plants (or of close, now extinct, polysporangiophyte relatives of the early vascular plants) with many homoiohydric features, i.e. an endohydric conducting system, cuticle, stomata and intercellular gas spaces (Taylor and Taylor, 1993; Remy et al., 1997).

The vascular plants (tracheophytes) are embryophytes; the other extant embryophytes are at the bryophyte grade of organization, and include the mosses, liverworts and hornworts (Gray, 1985; Kenrick and Crane, 1997). Here the sporophyte phase retains its early parasitism on the gametophyte phase throughout its life, so anchoring to, and absorption from, non-living substrata is essentially a gametophyte function in bryophytes.

While many aquatic, and some terrestrial, bryophytes are haptophytic, a majority are rhizophytic. The gametophyte phase of bryophytes often comprises multicellular structures on the land surface, with below-ground structures (rhizoids) comprising cylindrical evaginations of cells of the lower epidermis (sometimes with transverse walls). However, the Polytrichaceae and some other mosses have diageotropic rhizomes at or just below soil level as parts of axes continuing as leafy aerial shoots; the rhizomes bear rhizoids. Furthermore, the occurrence of below-ground positively gravitropic multicellular (parenchymatous) cylindrical branched structures in Takakia and Haplomitrium has been demonstrated (Grubb, 1970). While both genera were formerly regarded as liverworts, Takakia is now considered to be a moss (Smith and Davison, 1993; Kenrick and Crane, 1997). These ‘roots’ of bryophyte gametophytes lack a root cap, although the ‘roots’ become mucilage-covered in dry conditions (Grubb, 1970). Takakia and Haplomitrium ‘roots’ have a central cylinder of elongate cells, but, unlike the corresponding cells in the shoots of these plants, the central elongate ‘root’ cells are alive at maturity (Grubb, 1970) and so do not strictly function in the endohydric conducting system of these organisms. Grubb suggests that the superficial origin of branches in these ‘roots’ of bryophytes resembles that of the rhizospheres in the sporophyte phase of the vascular plant Selaginella, and that endogenous origin of branches is not a good criterion of the ‘root’ rather than ‘shoot’ nature of a structure (Grubb, 1970). The occurrence and roles of mycorrhizas in embryophytes of the bryophyte grade of evolution have been discussed; arbuscular mycorrhizas are common in the gametophytes of extant liverworts (Read et al., 2000).

Among the algae, the extant class of the Chlorophyta which is most closely related to ancestors of the Embryophyta is the Charophyceae. While probably not the nearest extant relatives of the embryophytes, the Characeae (Chara, Nitella, etc.) are rhizophytic aquatic plants which can have shoots up to 1 m long and have rhizoids penetrating several mm into sediments. These rhizoids are only one cell wide by one to several cells long, and are clearly involved in nutrient uptake (Raven, 1981, 1991). Many marine green macroalgae in the class Ulvophyceae, and some freshwater members of the Chlorophyceae, are also rhizophytes, usually with unicellular rhizoids. Most other benthic macroalgae are haptophytes. Many lichens are also, via the rhizines of the mycobiont, rhizophytes (Raven, 1981, 1986).

Julian et al. have used the term ‘root’ to denote a structure of the vestimentiferan Lamellibrachia which is embedded in anoxic sediments, and which seems to be involved in uptake of hydrogen sulphide by this animal with its chemolithotrophic symbionts (Julian et al., 1999).

These comparative analyses of the ‘rooting’ structures of rhizophytes (and quasi-rhizophytes such as Lamellibrachia) show that at least some of the classic root functions of the roots of sporophytes of terrestrial vascular plants are shared with the rhizoids or more complex structures of other phases in the tracheophyte life cycle, or other (non-tracheophyte) higher taxa of plants. The special role of rooting structures in water uptake is less significant in submerged aquatic plants where there is no transpiratory water loss, which accounts for up to 97% of the water taken up from the soil by C₃ plants on land. Furthermore, some specialized functions of roots demand the complexity of roots, or at least (parenchymatous) structures which are at least three cells in diameter. Examples are cluster roots (Marschner, 1995; Skene, 1998), epidermal transfer cells related to the external Fe³⁺ reduction mode of Fe acquisition found in almost all vascular plants tested (grasses are an exception) (Marschner, 1995), and the production of N₂-fixing nodules involving rhizobia or Frankia which need a minimal size for the required plumbing (carbohydrate input; organic N removal) and O₂-protective structures (Raven et al., 1994; Thumfort et al., 1994).

However, some important attributes of resource acquisition by higher plant roots can apply to morphologically simpler, and smaller-radius, absorptive structures, including rhizoids, root hairs and mycorrhizas. An example is apical alkalinization (increasing Mo availability and decreasing Al toxicity) and more basal acidification (favouring P and Fe availability) resulting from longitudinally and radially circulating currents carried by (buffered) H⁺ (Raven, 1989, 1991, 1995b, 1998, 2000a; Plassard et al., 1999). This could not directly apply to marine ulvophycean green algae with their currents carried by Cl^- (Raven, 1991). Furthermore, acidification of the whole rhizosphere by production of intracellular organic acids followed by K⁺/H⁺ exchange across the plasmalemma, or by organic acid (and, to a much smaller extent, amino acid) secretion across the plasmalemma, can occur even when NO₃^- is the N source and the expectation from N assimilation alone is for rhizosphere alkalinization (Bertrand et al., 1999; Darrah, 1993; Jones and Darrah, 1993; Jones et al., 1994; Kirk et al., 1994; Neumann and Römheld, 1999; Raven et al., 1980, 1990; Raven, 1995a, 2000a). Such acidification can aid P and Fe acquisition (Raven, 2000a). This organic acid excretion and H⁺ excretion (paralleling organic anion retention by roots) in more mature root zones accentuates the H⁺ efflux from circulating currents, just as alkalinization resulting from the organic anion efflux from root apices (related to offsetting Al toxicity) accentuates the alkalinization resulting from the H⁺ influx due to circulating currents carried by (buffered) H⁺. Kochian notes that alkalinization of the rhizosphere around the root apex does not invariably reduce Al toxicity (Kochian, 1995; cf. Raven, 2000a).

Finally, and as a link to the next section of this paper, the role of (aerial) shoots in nutrient acquisition (and release) must be mentioned. There has been a comprehensive review of land plant mineral nutrient acquisition from outside the pedosphere (Clarkson et al., 1986), including uptake of nutrient solutes by aerial shoots and other above-ground structures (Parsons et al., 1993, 1995). While not generally as important as the uptake of nutrients by the shoots of submerged rhizophytes, such uptake by land plants can be significant. This uptake by extant land plants often involves air-borne nutrients resulting from the activities of other land biota (or biota which evolved on land), for example, solutes leached from plant structures higher in the canopy in the case of understorey bryophytes, and uptake of gaseous NH₃ from seabird colonies or (anthropogenically) intensive animal production units (Raven and Yin, 1998). Accordingly, the earliest terrestrial biota would not have had access to the range and quantity of air-borne nutrients that are found today (Raven and Yin, 1998).

It is of interest that nutrient solutes which run off the canopy can induce rooting (Herwitz, 1991; Atger and Edelin, 1994).

	The biogeochemical context of the early embryophytes

Top Abstract Introduction What is a root? The functions of roots... The biogeochemical context of... The early (Palaeozoic) evolution... Implications for plants and... Conclusions References

 
Embryophytes evolved from charophycean green algae possibly as much as 510 million years ago (Gray, 1985; Edwards et al., 1998; Kenrick and Crane, 1997; Edwards, 2000; Strother, 2000; Wellman and Gray, 2000). The earliest (mid-Ordovician) fossil evidence as to the structure of the early embryophytes consists entirely of spores which were probably aerially dispersed (Gray, 1985; Edwards et al., 1998) and it appears that embryophytes evolved on land, possibly on the margins of drying pools, from relatively small algal ancestors (Stebbins and Hill, 1980; Graham, 1993).

The land surface inhabited by the earliest embryophytes was one that was already inhabited by photosynthetic organisms and by organisms which consumed, pre- and post-mortem, these photolithotrophs. There is geochemical evidence consistent with organisms living on land back to 1.2 billion years ago, i.e. in the Meso-Proterozoic (Horodyski and Knauth, 1994; cf. Keller and Wood, 1993; Shear, 1991; see also Rye and Holland, 2000). The argument here is based on the weathering of rock surfaces. The rate of weathering depends on temperature, the presence of water, and the concentration of CO₂ in this rock-surface water. The rate of weathering on land surfaces between 800 and 1200 million years old is said (Horodyski and Knauth, 1994), to have exceeded the rate expected from the likely atmospheric CO₂ concentration equilibrated with surface water, with the atmospheric CO₂ estimated from the greenhouse considerations and the lower radiation output from the sun 800–12 000 million years ago than is the case today. While there are no fossils of the actual organisms involved, evidence from marine deposits shows that cyanobacteria and (prasinophycean) green algae were present in the sea, and could have been primary producers on land as well (Raven, 1995a, 1997, 1998; Golubic and Seong-Joo, 1999). Parasites, grazers and decomposers could have been viruses, bacteria, archea, and (possibly) fungi; Raven (1997).

The biological ‘CO₂ pump’ from the atmosphere to water in fissures in the rock surface would have taken the form of primary producers (cyanobacteria; green algae?) on the rock surface where light supply was sufficient to allow photosynthesis. Transport of organic material into the fissures on the rock could have occurred as particles (dead cells gravitationally falling into crevices, motile protists, or fungal filaments, moving into the crevices; and soluble organic material released from dead, including virally lysed, cells percolating into crevices). Subsequent metabolism, with respiratory or fermentative CO₂-production, could have led to a CO₂ build-up as diffusive CO₂ movement to the atmosphere only equals net CO₂ production in the rock crevices when a significant concentration difference is established between the sites of CO₂ generation and the water-atmosphere interface. This build-up of CO₂ means that CO₂ consumption in rock weathering becomes an increasingly significant competitor with diffusive CO₂ loss to the atmosphere. The impact of cyanobacteria and algae in the weathering of mineral surfaces today has been discussed (Vestal, 1993; Walton, 1993; Wynn-Williams, 1993).

The chemistry by which CO₂ is removed in weathering of silicates can be represented as (equation (1)):

(1)

The HCO₃^-, Si(OH)₄, Ca²⁺, and Mg²⁺ dissolve in the soil water which becomes groundwater, lakes and rivers and ultimately reaches the sea. Here the Si(OH)₄ is precipitated (with much biological participation) as SiO₂, and Mg²⁺ and (especially) Ca²⁺ is precipitated as MgCa(CO₃)₂ (dolomite) and CaCO₃ (calcite) according to equations (2 –4):

(2)

(3)

(4)

Overall, equations (1)+(2)+(3) sequester 2CO₂ (ultimately from the atmosphere) as 2 sedimented .

Equations (1–4) occur over time scales of millions to tens of millions of years. The exogenic C cycle is completed by magmatic and volcanic processes in the Earth's crust. Here CaMg(CO₃)₂, CaCO₃ and SiO₂ produced in equations (2–4) are converted to the starting points (other than H₂O) for equation (1) as follows:

(5)

This part of the cycle takes of the order of hundreds of millions of years.

Any CaCO₃ which reaches the surface of the land is weathered according to equation (6):

(6)

This, with equation (4), shows that CaCO₃ weathering on land and production in the sea is CO₂-neutral over times of tens to hundreds of millions of years. The time scale here relates to the predominant deposition of CaCO₃ in continental shelf sediments, i.e. above the lysocline. It also concerns the long time taken for shelf CaCO₃ to be weathered; the shelf deposits are generally not subducted and subject to equation (5) reactions, but follow equation (6).

Weathering of rocks of varying mineralogy has a very wide range of chemical effects (Berner and Berner, 1996), but it generally increases the solubility of many of the plant nutrients found in rocks, for example, potassium, magnesium, calcium, phosphate, and sulphate among the macronutrients. The atmosphere 1200 million years ago contained oxygen (all from photosynthesis in the sea) at perhaps 1% or more of the present level, and would have immobilized the micronutrient iron as ferric oxides, hydroxides and carbonates; the ferric oxides, with aluminium oxide also produced by weathering, immobilize phosphate. Iron and phosphate could be remobilized by organic acids and reductants produced by biota and, in the case of iron, by siderophores produced by many bacteria (including cyanobacteria) (Raven, 1997). In addition to sources from weathering on land, sulphur could have been supplied via the atmosphere from oxidation products of dimethylsulphide produced by marine biota (Lovelock et al., 1972).

Nitrogen is a plant nutrient present in very small amounts from rock weathering relative to other nutrient elements when viewed in the context of the ratio of elements found in optimally nourished plants. The large pool of N₂ in the atmosphere can only be accessed biologically by prokaryotes; there is the possibility that bacteria (including cyanobacteria) and archea could have been fixing N₂ on land in the Proterozoic (Raven and Yin, 1998; Golubic and Seong-Joo, 1999). The co-occurrence of N₂ and O₂ in the Meso- and Neo-Proterozoic atmosphere permitted the generation of biologically significant quantities of oxidized combined nitrogen in electrical storms (Raven and Yin, 1998).

Work on extant terrestrial cyanobacteria, algae and chemo-organotrophic microbes on rocks (Vestal, 1993; Walton, 1993; Wynn-Williams, 1993) suggests that their influence on weathering is less than that caused by lichens. However, lichens were not present in the Proterozoic; the earliest known fossil lichen is from the Rhynie Chert (Lower Devonian) in the form of a lichen with a cyanobacterial photobiont and a glomalean mycobiont with no present-day analogues (Taylor et al., 1997). The ascomycetes and basidiomycetes, which are the fungal components of extant lichens, are not known as fossils until the Upper Silurian (based on microfossils: Sherwood-Pike and Gray, 1985; Taylor, 1994) and the Carboniferous-Permian, respectively, although molecular–genetic evidence, combined with zygomycete fossils, suggests an origin of filamentous fungi 600 million years ago (Simon et al., 1993; Simon, 1996; Blackwell, 2000; Redecker et al., 2000). The present form of lichens probably evolved in the Permo-Triassic as peltigeralean lichens (Hawksworth, 1988) although the earliest known fossils of lichens involving fungi other than zygomycetes are only 50 million years old.

These considerations suggest that a skeletal soil could have been present as much as 1.2 billion years ago. The presence of O₂ in the atmosphere at this time would have had direct (metabolism, including respiration and active oxygen species production) and indirect (UV-B screening via stratospheric O₃, and restrictions on P and Fe availability) effects on land biota.

The authors do not wish to over-emphasize the extent to which the marine biota and the microbial terrestrial biota had modified the terrestrial habitat in ways which might have facilitated the establishment of a rhizophytic terrestrial embryophyte flora. However, it is likely that the land surface was preconditioned to some extent as a habitat which the early embryophytes could further modify upon their evolution perhaps 510 million years ago. The use of geochemical proxies (in this case the Fe(CO₃)OH content of goethite) to estimate the rock surface (‘soil’) CO₂ concentration (Yapp and Poths, 1992, 1994) on land 440 million years ago is consistent with high primary productivity, which at once permits (via CO₂ input) and requires (to support productivity) high rates of weathering or, at least, of recycling of nutrients from biota. The high productivities suggested (Yapp and Poths, 1994) requires quite complex plant structures (Raven, 1997, 1998), for which the only fossil evidence is the aerially dispersed spores (Gray, 1985). Suggestions that spores of potentially mycorrhizal glomalean fungi occur in Ordovician deposits (Blackwell, 2000; Redecker et al., 2000) require reconsideration in view of the mineralogical context.

A further point about the high productivities suggested (Yapp and Poths, 1994) in the Ordovician concerns the impact on atmospheric CO₂. Increased weathering rates tend to draw down atmospheric CO₂, depending on the land area covered by these high-productivity plants. This has implications for the later discussion on the role of deeper rooting in the Devonian CO₂ drawdown shown by the Geocarb II model of Berner (Berner, 1994), since one of the inputs to the model is the plant-induced CO₂ drawdown (Raven, 1998).

	The early (Palaeozoic) evolution of embryophytes and the origin of roots

Top Abstract Introduction What is a root? The functions of roots... The biogeochemical context of... The early (Palaeozoic) evolution... Implications for plants and... Conclusions References

 
Silurian
Edwards et al. (Edwards, 1993, 1994; Edwards et al., 1998) discuss the evidence as to the vegetative morphology and anatomy of the earliest embryophytes known from remains other than spores (Banks, 1970; Chaloner, 1970; Spicer, 1989; Kenrick and Crane, 1997; Bateman et al., 1998; Shear, 1991). These Silurian fossils are generally composed of branching parenchymatous axes, and appear to represent the sporophyte phase of the life cycle, based on the occurrence of sporangia or, where these are not present, the absence of antheridia and archegonia, or of a structure representing a young sporophyte parasitic on a gametophyte, as is always the case for bryophytes and homosporous free-sporing tracheophytes. That these two latter indicators of the gametophyte nature of a fossil can be preserved is shown by Lower Devonian fossils with well-preserved antheridia and archegonia or vegetative structures which are more complex than any extant gametophyte (see below) (Taylor and Taylor, 1993), and a liverwort-like fossil with gametophyte and (possibly) an attached sporophyte (Edwards et al., 1995).

The clearly sporophytic (i.e. sporangia-bearing) fossils from the Silurian have, in many of the cases where preservation is adequate to show the structures if indeed they were present, one or more of the four essential attributes of homoiohydry, i.e. an endohydric conducting system, cuticle, stomata, and some underground structures. Preservation of Silurian fossils examined so far is never adequate to show if the fifth attribute of homoiohydry, i.e. intercellular gas spaces, occurred. The full suite of five structures permits water (and nutrient solute) supply to aerial axes when enough soil water is available to supply the transpiring surface with liquid water at a rate sufficient to support the transpiratory loss of water vapour found during rapid photon and CO₂ absorption in photosynthesis. By stomatal closure organisms with these attributes can maintain cellular hydration for shorter or longer periods (depending on, for example, the water conductance of the cuticle plus epicuticular wax) when soil water supply is inadequate to make good the transpiratory water vapour loss when stomata are open. Stomatal closure prevents net CO₂ fixation, and thus trades death from thirst for the potential of death by starvation, but by tiding the plant over periods of limited soil water supply as a hydrated vegetative entity the plant can grow on land and yet not be tolerant of desiccation in the vegetative state. For as yet imperfectly understood reasons the combination of homoiohydry and vegetative desiccation intolerance seems to be essential for any terrestrial plant which has structures more than about a metre above ground level (Raven, 1993, 1995a, 1997, 1998).

It must be emphasized that in most of these Silurian sporophytes the occurrence of all of the attributes of homoiohydry in a given (form) genus or species is very much the exception rather than the rule. Even for one of the most complex known (Upper) Silurian sporophytes, the (probable) lycophyte Baragwanathia, cuticle, stomata, tracheids, and (probably) intercellular gas spaces are known from above-ground structures in Lower Devonian specimens (but not so far from the less well preserved Upper Silurian fossils) (Hueber, 1983), but we have little knowledge of the underground structures which, it is widely believed, such plants must have possessed if they were to have been adequately anchored, and to have had a sufficient capacity for uptake of water and soil-derived nutrients (Driese and Mora, 2001). The aerial shoot of Baragwanathia had microphyllous leaves, with sporangia whose mode of attachment (to the upper leaf surface or to the axis just above leaf insertion with no evidence of any consistent relation between microphyll and sporangium) is still enigmatic. Any below-ground axes of Baragwanathia would be expected to lack the stomata, microphylls and sporangia as well as the overt cuticle. Without these features, and with generally imperfect preservation of the anatomy of Baragwanathia (at least in comparison with the Rhynie Chert fossils), how would such an axis system be recognized as being from Baragwanathia? It is clear from the size of the erect aerial axes of Baragwanathia that substantial below-ground structures would be needed.

Lower Devonian
In the Lower Devonian there are fossil localities with exceptional preservation, for example, the Rhynie Chert, and also localities with important evidence of the earliest rooting structures (Driese and Mora, 2001; Gensel et al., 2000; Hotton et al., 2001). Here a range of embryophyte sporophytes and gametophytes is found. The sporophytes have branched aerial axes (i.e. are polysporangiophytes: Kenrick and Crane, 1997) which are endohydric, but do not always have true xylem (i.e. water conducting cells with unevenly but regularly thickened walls typical of tracheids); an example of an endohydric plant without true xylem is Nothia aphylla (Kerp et al., 2001). Regardless of whether they have true xylem, these sporophyte axes have intercellular spaces, cuticle and stomata, and sporangia. Among those with true xylem are representatives with smooth axes (Rhynia gwynne-vaughanii) and, with microphylls, the lycophyte sensu lato Asteroxylon mackiei. Aglaophyton (and some other Rhynie plants) have underground axes bearing rhizoids; these axes have arbuscular mycorrhizas, but mycorrhizas also occur on above-ground axes (Nicolson, 1975; Pirozynski and Malloch, 1975; Pirozynski and Dalpé, 1989; Simon et al., 1993; Taylor et al., 1995; Simon, 1996; Smith and Read, 1997). The occurrence of rhizoids may not always be an indicator of the subterranean nature of axes, since rhizoids of Rhynia are borne on hemispherical projections near the base of presumed aerial axes (Edwards, 1993). Edwards also points out that Asteroxylon sporophytes have anatomical differences (e.g. in the shape of the stele) between leafy aerial and smooth subterranean axes, but that distinction of the smooth axes as roots is not possible on the basis of endogenous branching, since the frequent branching by the smooth axes is always exogenous (Edwards, 1993). In the single known example of an Asteroxylon leafy shoot with endogenous branching it is not clear whether the branch is leafy or smooth (Edwards, 1993). In any case there is Grubb's warning about the use of endogenous branching as an infallible test of the root nature of an organ (Grubb, 1970), although his tracheophyte example, i.e. the extant Selaginella, is perhaps contentious (Bierhorst, 1971; Gifford and Foster, 1987; Lu and Jernstedt, 1996).

Fossils of other Lower Devonian embryophyte sporophytes, i.e. those from sites other than the Rhynie Chert, also have strong indications of the occurrence of roots, or at least structures with considerable similarity to roots. Thus, Rayner (Rayner, 1984) has suggested that the lycophyte (sensu lato) Drepanophycus spinaeformis had roots on the basis of the branching pattern and the attitude in the sediment of the smooth axes in relation to the bedding plane (Schweitzer, 1980). However, there is no anatomical evidence (e.g. endogenous branching) to support this view. Li and Edwards have re-investigated Drepanophycus spinaeformis from the Lower Devonian of Yunnan Province, southern China (Li and Edwards, 1995). This material, now named Drepanophycus qujingensis (Li and Edwards, 1995), was shown to have roots, i.e. structures borne by both fertile and sterile leafy shoots, which branch dichotomously at least five times. The roots appear to be inserted randomly, and the largest branching root system extends up to 30 mm from the stem. The diameter of the roots is about 3 mm at their attachment and rapidly decrease with branching; the smallest branches recovered have a diameter of less than 0.7 mm (Fig. 37D of Li and Edwards, 1995), and it is likely that the unpreserved, more apical, portions had an even smaller diameter. It is of interest that these roots were borne on leafy axes which were presumably above-ground but were probably prostrate, as the rooting system itself extends down into the rock, at an angle to the bedding plane in which the leafy axes were exposed.

The published information on roots or root-like structures in the early Devonian (Pragian-Emsian) (Gensel et al., 2001) has been reviewed, and new data presented on Lower Devonian fossils from Bathurst Island, Arctic Canada, and from New Brunswick and Gaspé, Canada. These plants with root-like structures are zosterophyllophytes or lycophytes sensu lato. Rooting structures in these fossils were identified by the following five characteristics granted that the aerial axes that bear the rooting structures have emergences or microphylls (Gensel et al., 2001). (1) The rooting structures are attached to aerial ones, but have no emergences or microphylls. (2) The rooting structures are generally much narrower in diameter, and shorter, than are the aerial axes that bear them. (3) The rooting structures are either unbranched or are irregularly branched; this contrasts with the more predictable branching of the aerial axes. (4) The rooting structures may have a sinuous or delicate appearance. (5) The rooting structures proceed in a direction opposite to that in which the aerial axes grow. These authors point out that characteristics 2–5 are also applicable to plants with naked axes, and that the rooting structures show no evidence of root caps or indisputable root hairs; furthermore, the anatomy of the rooting structures is not yet known.

The Bathurst Island fossils (Gensel et al., 2000) comprise Drepanophycus and Bathurstia spp. with aerial axis-borne rooting structures scattered along what appear to be trailing or rhizomatous structures, or forming parts of the so-called ‘K-branching’ structures, and Zosterophyllum sp. nov. with tufted rooting structures which occur as descending naked axes from plants with dense aggregations of branching aerial shoot systems. The Quebec and New Brunswick fossils comprise Crenaticaulis verruculosus and Sawdonia ornata: C. verruculosus has smooth, spinous, slender (0.5 mm wide) rooting structures departing at right angles to aerial or rhizomatous axes in regions where there are otherwise no branches (Gensel et al., 2001). Sawdonia ornata fossils have structures like the putative rooting structures of C. verruculosus in close association to aerial axes, but not attached to the S. ornata aerial axes. The association of the root-like structures with S. ornata is rendered more likely because S. ornata is essentially the only species present at that level in the bedding sequence. Furthermore, Rayner interprets structures associated with H-shaped branching Scottish Emsian specimens of Sawdonia ornata as root-like or as rhizophores (Rayner, 1983). These S. ornata specimens have one of the usually basipetally directed, smooth laterals produced at H-shaped branching which differs from the usual spiny branches; sometimes there is a short spineless projection in the position normally occupied by this longer structure (Rayner, 1983).

The discussion above shows that the evidence for roots in Lower Devonian fossil sporophytes is based on the occurrence of smooth subterranean axes with branching patterns similar to extant roots, and with a simpler xylem morphology (protostele) than the actinostele in aerial parts of plants such as Asteroxylon. Where the origin of branches in these underground axes is known, it is superficial. There is no evidence of a root cap, although this is not a commonly fossilized structure; an exception is the well-preserved root cap of the Pennsylvanian (Upper Carboniferous) marattacious fern Psaronius (Stewart and Rothwell, 1993). Furthermore, these smooth axes apparently lack an endodermis (which can be found in Carboniferous and later vascular plants, for example, in roots of Equisetites, Tempskya and Rhizopalmoxylan: Walton, 1940; Raven, 1984b, 1994b; Tidwell, 1998). However, there is no evidence of an endodermis in aerial axes of those Lower Devonian plants, and the extant Lycopodium sporophytes also lack a root (but not a shoot) endodermis (Damus et al., 1997; Raven, 1983, 1984a, b, 2000a, b; Edwards et al., 1996; Epstein, 1999), as do leafy endohydric gametophytes of extant bryophytes (Grubb, 1961).

It would seem that the Lower Devonian sporophytes had below-ground parenchymatous structures which performed the functions of roots, but which do not necessarily have all of the features by which the roots of extant vascular plants are recognized. Furthermore, the zosterophyllophyte-lycophyte sensu lato clade possessed root-like structures in the Lower Devonian, although there is no corresponding evidence for root-like structures on euphyllophytes at this time. The earliest convincing euphyllophyte roots occur in the Middle Devonian cladoxylalean Lorophyton goense, thought to derive from trimerophytes (Fairon-Demart and Li, 1993) where bifurcating structures arise in tufts from a swollen stem base. This is consistent with the polyphyletic origin of roots, a point considered in more detail below.

Upper Devonian and Carboniferous
In Upper Devonian, and, especially, Carboniferous deposits, root structures with a protostele, endogenous branching and an endodermis can be distinguished from shoots of the same plant with (often) a more complex anatomy of the stele in primary growth, exogenous branching and (often) an endodermis (Taylor and Taylor, 1993; Meyer-Berthaud et al., 1999). The most parsimonious explanation of these data is that endodermes, like roots, are not monophyletic in origin. The same conclusion applies to the origin of the exodermis, which is found in 90% of the 200 extant angiosperm species so far examined, but in none of the extant gymnosperms examined, and in only a few of the extant non-seed vascular plants investigated (Damus et al., 1997).

An interesting case is the arborescent lycophytes. The embryo of the Upper Carboniferous lepidodendrid Lepidocarpon produces a ‘root-like’ organ by dichotomy of the embryonic shoot; this recalls the way in which root-like structures of Lower Devonian zosterophyllophytes and lycophytes sensu lato originated (Gensel et al., 2001). This root-like organ develops into the ‘stigmarian root system’ bearing ‘stigmarian rootlets’, and the stigmarian axis and its rootlets is considered homologous to the shoot system (Rothwell, 1984; Rothwell and Erwin, 1985). This view is supported by the presence of a pith cavity in the stigmarian root system, and by the exogenous production of lateral organs (stigmarian rootlets) (Stewart and Rothwell, 1993; Taylor and Taylor, 1993). The ‘stigmarian rootlets’ have thus been interpreted as morphologically equivalent to leaves, and are closely similar in anatomy to the roots of the extant Isoetes spp. where, as one might expect of modified leaves, the roots have a major role in CO₂ uptake by the plant (Raven, 1970; Sand-Jensen et al., 1982; Richardson et al., 1984; Keeley et al., 1984; Raven et al., 1988). However, such an origin of roots is not a prerequisite for a major role in CO₂ uptake, for example, in flowering plants of the isoetid life form (Raven et al., 1988). The absence of root hairs on ‘stigmarian rootlets’ is also consistent with their origin from leaves.

Fossil and other evidence bearing on monophyly or polyphyly of roots
Aside from the ‘stigmarian rootlets’, it is necessary to address the question of the occurrence of a single evolutionary origin as opposed to multiple origins of roots (Groff and Kaplan, 1988). The question can be posed as the difference between a single origin before lycophytes and euphyllophytes diverged, and independent origins in each of these two clades.

The lycophyte clade (in the form of the putative lycophyte Baragwanathia) can be seen as separate from other tracheophytes by the Late Silurian (400–420 million years ago), with the rhyniophytes related more closely to the ancestors of the extinct trimerophytes and other fossil and extant euphyllophytes which include all extant vascular plants other than lycophytes (Kenrick and Crane, 1997; Dolan and Scheres, 1998). The presence of clearly defined root-like structures in fossils of lycophytes in the Lower Devonian (see above), but their absence from equally well-preserved Lower Devonian fossils of other vascular plants (e.g. in the Rhynie Chert) suggests independent origins of roots in the lycophytes and the euphyllophytes. The alternative explanation is that roots existed in both clades in the Lower Devonian and the Upper Silurian, and in their common tracheophytic ancestor and that the absence of roots from the late Silurian fossil record, and in non-lycophytes in the Lower Devonian is a result of poor preservation of roots relative to other axial structures, including sporangia. This is possible for Late Silurian specimens of Baragwanathia (see Edwards, 1993, 1994 for discussions on the preservation potential of roots lacking secondary thickening), but is perhaps unlikely as a general explanation. The first possibility, i.e. an independent origin of roots in lycophytes and (later) in euphyllophytes, requires that such features of roots as a root cap, and endogenous origin of laterals evolved independently from the original relatively undifferentiated axis of the lycophytes and the euphyllophytes. Certainly the selective advantages of a root cap in an organ growing through soil seem clear and could account for two independent evolutionary origins. The endodermis is also found in many extant shoots (especially among organisms of the pteridophyte grade of organization) as well as in almost all extant roots (Lycopodium spp. being an exception). As for the simpler and less diverse morphology of the steles of roots (medullated or non-medullated protostele or actinostele) than of shoots (all categories of stele morphology occur in shoots), the root situation is closer to that of the ancestral axis.

A further line of evidence comes from the embryology of extant vascular plants. Goebel distinguishes allorhizic plants, where shoot and roots are at opposite ends of the embryonic axis, from homorhizic plants, where the root is lateral with respect to the longitudinal axis of the embryo (Goebel, 1930). All extant seed plants are allorhizic, while all extant free-sporing vascular plants are homorhizic. If this is a fundamental evolutionary distinction, then extant vascular plants (and their progymnosperm and (?) trimerophyte ancestors) can be distinguished from the lycophytes and extant euphyllophytes (Equisetum; ferns). This suggests the possibility of three origins of roots, yielding the allorhizic seed plants and their root-bearing vascular ancestors, the homorhizic lycophytes, and the homorhizic euphyllophytes other than seed plants (e.g. ferns and horsetails) and any of their ancestors from which they inherited roots.

A further, and much more recent, approach to the origin of roots comes from molecular genetic studies of shoot and root development in the (allorhizic) flowering plant Arabidopsis thaliana (Dolan and Scheres, 1998). These authors investigated the propagation of root pattern, formed in the embryo, through subsequent growth and development of the seedling root meristem. The root pattern is maintained by positional cues, and to some extent by local cell interactions. These cell interactions are involved in maintaining the balance between cell division and differentiation in cells close to the quiescent centre. This balance is also maintained in shoots, where the genetic basis of the phenomenon is now understood. They also point out that patterned cell differentiation in the epidermis of roots has a similar genetic basis to that in shoots (Dolan and Scheres, 1998).

Dolan and Scheres point out that the developmental similarities of shoots and roots in Arabidopsis could indicate that roots are derived from whole shoot structures (Dolan and Scheres, 1998). An alternative possibility (Dolan and Scheres, 1998) is that a suite of genes involved in shoot structure determination could have been co-opted into a similar role in roots derived from a lateral appendage of the shoot (other than a branch shoot), such as is likely to be the case for at least some (homorhyzic) lycophytes, i.e. the isoetalean clade with roots derived from the leaf-like structures of stigmarian rootlets. The stigmarian roots on which the stigmarian rootlets are borne are derived by dichotomy of a shoot (see above). Dolan and Scheres acknowledge that this sort of molecular genetic work should be extended to other vascular plants, including free-sporing plants (Dolan and Scheres, 1998).

It was concluded from these lines of evidence that roots are probably (at least) diphyletic. Further genetic approaches to root morphogenesis in lycophytes and euphyllophytes should give a more definitive answer.

The diameter of the earliest roots
In view of the functional significance of root diameter in extant plants (see above) it is important to know the diameter of the earliest roots, both in the fossil record and in each major grade of embryophytes.

The earliest roots in the fossil record have diameters ranging from 3 mm at their attachment to less than 0.7 mm in the smallest branches recovered (Li and Edwards, 1995; see above). It is possible that roots of even smaller diameter occurred in these Lower Devonian lycophytes, but have not been preserved and recovered; the fossil record is likely to be biased in favour of the larger rather than the smaller diameter axes of roots lacking secondary thickening (Li and Edwards, 1995).

For extant embryophytes, the data for free-sporing polysporangiophytes suggest that the smallest lycophyte roots can be 330 µm (Lycopodium)–370 µm (Isoetes) in diameter, while among the euphyllophytes Equisetum has a root diameter of 370 µm or more, while fern roots range from 140 µm (Azolla and Hymenophyllum) to 830 µm (Ophioglossum) (Smith, 1938; Schneider, 1996). Azolla and Hymenophyllum are clearly relatively derived ferns (Rothwell, 1996). All known extant and extinct gymnosperms have secondary thickening and are mycorrhizal, although the finest branches of roots without secondary thickening have diameters of only a few hundred µm (Smith and Read, 1997). Among flowering plants, the roots of those considered, on molecular phylogenetic grounds, to be the closest living relatives of the ancestral angiosperms have roots of relatively large diameter (Baylis, 1975; St John, 1980; Qui et al., 1999; Soltis et al., 1999). Such roots may have low elongation rates and (except in aquatics such as the Nymphaeales) have a great dependence on mycorrhizas.

	Implications for plants and for their environment of increased rooting depth by vascular plants during the Palaeozoic

Top Abstract Introduction What is a root? The functions of roots... The biogeochemical context of... The early (Palaeozoic) evolution... Implications for plants and... Conclusions References

 
The discussion of the biogeochemical content of the early embryophytes suggested that pre-embryophytic photosynthetic organisms on land, and associated microbial food webs, had significant effects in increasing the rate of weathering and the mobilization of plant nutrients from rocks, and in N₂ fixation, leading to the production of a surface layer resembling soil. The evolution of embryophytes with roots (or with the antecedents, or functional analogues, of roots) increased this weathering activity. The larger and more structurally complex below-ground structures, involved in anchoring the plant and in taking up water and nutrients, were simultaneously permitted and required by larger and more structurally complex above-ground structures absorbing photons and CO₂.

Chaloner and Sheerin show, in their text—Fig. 5 (see also Fig. 6.15 of Niklas, 1997) the maximum observed aerial axis diameter of vascular plant fossils from the late Silurian to the Devonian-Carboniferous boundary (Chaloner and Sheerin, 1979). The increase is from some 3 mm diameter in the latest Silurian, via an approximately linear increase in the logarithm of diameter with time, to almost 2 m at the end of the Devonian in the progymnosperm Archaeopteris (Callixylon). This progymnosperm has secondary thickening; the use of the relationships between height and basal stem diameter for organisms with secondary thickening (Fig. 5.9, Niklas, 1994; Figs 6.14, 6.15, Niklas, 1997) suggest a height for Archaeopteris of 10–30 m.

Corresponding to this increase in height is an increased plant biomass per unit land area and depth of penetration of roots sensu lato. While not suggesting that the depth and mass of below-ground parts of the plant increase in direct proportion to the height and mass of above-ground parts of the plant, there is independent evidence of an increasing depth of penetration of roots (Algeo et al., 1995; Algeo and Scheckler, 1998; Berner, 1990, 1993, 1994, 1997, 1998; Retallack, 1997; Elick et al., 1998) during the Devonian. Thus, dichotomous root traces up to 2 mm in diameter and up to 0.9 m long in the Emsian (late Early Devonian) have been found (Elick et al., 1998). An increasing area of land occupied by tracheophyte vegetation, and an increasing rooting depth in vegetated areas, meant an increasing volume of rock weathered throughout the Devonian. This permitted the ‘mining’ of more rock-derived nutrients which supplied the increasing biomass of plants, although the need per unit of added biomass for P, K, Fe and Mg does not increase in proportion to biomass in plants with extensive secondary thickening, since much of the biomass comprises non-living xylem which has much lower P, K, Fe and Mg (and N) content relative to the peripheral living tissue. This argument on mineral requirements requires that the P, K, Fe, Mg and N released can be recycled to growing peripheral tissue. A similar consideration applies to the recycling of P, K, Fe, Mg and N from relatively short-lived photosynthetic structures and nutrient and water-acquiring below-ground structures in long-lived plants (Addicott, 1982; Atkinson, 1992; Raven, 1986; Jackson et al., 1997; Robinson, 1990). Internal recycling from senescent short-lived structures into the perennial plant structures avoids the loss of these resources which have not been transported into the living core of the plant when the short-lived organ dies, regardless of whether it is abscissed. There are also considerations of how rapidly these mineral elements are recycled from dead plants (or abscissed or otherwise removed dead parts of plants) to the soil and hence to growing plants.

It has been pointed out that the increased weathering rate, as the volume of land surface impacted by root systems increased during the Devonian, was probably transient (Algeo and Scheckler, 1998). The argument contrasts to the high atmospheric CO₂ with minimal (see above) biological pumping of CO₂ into the land surface before the deep ‘rooting’ of embryophytes in the Devonian radiation with the low atmospheric CO₂ (due to increased weathering on land) with higher CO₂ in the land surface due to increased biological pumping as the volume of land surface impacted by root systems from the Carboniferous onwards. These two situations yield essentially identical rates of weathering (averaged over millions of years) before and after the Devonian radiation (Algeo and Scheckler, 1998). To recapitulate, the pre-Devonian high atmospheric CO₂ and minimal biological CO₂ pumping gave way to a post-Devonian lower CO₂ atmospheric level with substantial biological CO₂ pumping into the land surface and both scenarios are suggested to yield similar weathering rates. In the Devonian, as the area, and even more, the volume of biologically weathered land surface increased very significantly. This rapid weathering removed CO₂ from the atmosphere and was maintained by the increased volume of land surface impacted by high CO₂ produced from the biological pump. It has been suggested that CO₂ in the land surface built up relative to the (decreasing) atmospheric CO₂ in the Devonian so that the weathering rate was higher than before or after (Algeo and Scheckler, 1998). The argument of these authors suggests that there was a pulse of weathering of the land surface in the Devonian, whose extent has not been matched before (when weathering was predominantly a result of high atmospheric CO₂) or since (when weathering was predominantly a result of high land surface CO₂) (Algeo and Scheckler, 1998). It is possible that Algeo and Scheckler somewhat underestimated the pre-Devonian role of non-vascular terrestrial photosynthetic and other organisms in enhancing weathering (Algeo and Scheckler, 1998; Horodyski and Knauth, 1994; Yapp and Poths, 1992).

A further effect of increased weathering and the production of soil concerns the increased storage capacity for water near the surface. This helps to tide homoiohydric plants over between stochastic rainfall events as photosynthesizing organisms rather than just permitting them to stay alive, as would be the case for poikilohydric, desiccation-tolerant organisms. This water storage, together with the moderating effect on rainfall of the above-ground plant canopy and the more extensive root system, can reduce soil erosion during rainstorms.

These effects of the increased rock weathering as a result of an increased volume of rock available to weathering by plants have been viewed until now as having essentially a positive feedback in plants, i.e. increasing the quantity and/or continuity of resource (water, pedosphere-derived nutrients) availability. However, plant-increased weathering also has an effect on the aerial environment of plants (discussed above in relation to the transitory increase in weathering rate) which can be considered as a negative feedback on the capacity for the rate of biomass increase. This relates to the removal of atmospheric CO₂ in weathering of silicate minerals.

The mechanism whereby photosynthesis at the (rock or) soil surface and respiration below the photic zone (a concept borrowed from aquatic ecology) increase CO₂ use in weathering, which was outlined earlier for algal mats as part of microbial food webs, is quantitatively accentuated for tracheophytes, and especially for deeper-rooting tracheophytes. The increased external (via a leaf area index in excess of one) and internal (via intercellular gas spaces in leaves) area involved in CO₂ uptake permits the canopy-top photon flux density of up to 2000 µmol photon m^-2 s^-1 to be used effectively in photosynthesis, noting that the potential for CO₂ fixation could be constrained with less appropriate geometry by diffusion of CO₂ in solution to the carboxylase. This latter constraint might, with geometry of photosynthetic structures similar, in principle, to those of extant plants, in the Silurian and early Devonian with CO₂ at 10–20 times the present level, permit C₃ photosynthesis to occur at up to 6 µmol CO₂ per m² of area over which CO₂ passes from the gas to the solution phase per second. This, with a photon yield of C₃ photosynthesis in high CO₂ of 0.1 mol CO₂ fixed per mol photons absorbed or 0.08 mol CO₂ fixed per mol incident photons (assuming an absorptance of 0.8), would permit a fixation of up to 75 µmol CO₂ m^-2 gas exchange surface s^-1; this requires a gas-exchange surface area of 27 times the land area to effectively use 2000 µmol photon m^-2 s^-1 incident photon flux density. This could be accomplished by a leaf (axis) area index of three and an internal area exposed to a gas phase nine times the external (projected) area (see Raven, 1977, 1993, 1995a, 1997, 1998; Konrad et al., 2000).

Up to half of this productivity could be transported in phloem (or analogous) tissues to roots, where up to half of the transported organic carbon could be respired to CO₂ by living plant cells. Dead roots, plus aerial plants parts mixed into soil by animals, could be respired by animals and, especially, microbes to produce CO₂. This CO₂ has two possible sinks. One sink is diffusion from the soil solution, via the soil gas phase to the atmosphere. The other sink is consumption in weathering. The low conductance for CO₂ via the soil gas spaces resulting from the long linear diffusion path and tortuosity of the diffusion path, and the small fraction of the area parallel to the soil surface which is occupied by gas spaces, result in the build-up of CO₂ in the rock/soil surface layers. This CO₂ build-up increases the rate of weathering. The balance between CO₂ evasion to the atmosphere and CO₂ consumption in weathering as CO₂ output into the rock/soil surface layers involves an increased rate of weathering with an increased rate of respiratory CO₂ injection into the rock/soil surface. It is of interest that CO₂ arises not only from respiration by below-ground plant organs, and by heterotrophs living on live or dead particulate plant material, but also from respiration of soluble organic solution. These organic acids can, before their conversion to inorganic carbon, increase the rate and pathway of weathering of silicates (Berner and Berner, 1996).

The occurrence of an increasing fraction of the area of the land surface colonized by tracheophytes, and an increasing depth of (rock) soil exploited by tracheophyte roots, tends to speed up the reactions in equation (1) relative to those of equation (3), but has a negligible effect on the rate of equation (5) (see above under The biogeochemical context of early embryophytes). This means that the expansion of tracheophytes, including the activities of their root systems, removes CO₂ from the atmosphere. While the build-up of organic matter in biomass also removes CO₂ from the atmosphere, the much more rapid turnover of the great majority of this material to regenerate CO₂ (months to thousands of years, depending on the chemistry and structure of the material: Schlesinger, 1991; Berner and Berner, 1996) means that the impact on atmospheric CO₂ over millions of years and longer comes more from the sedimentation, weathering and vulcanism of inorganic carbon than of organic carbon.

The increased weathering due to more widespread and deeper-rooting tracheophytes is thought to be a major cause of the decrease in atmospheric CO₂ during the Devonian to a minimum (until the recent Pleistocene glacial maxima) value in Carboniferous (Berner, 1990, 1993, 1994, 1997, 1998; Berner and Berner, 1996). However, it is noted that the enhanced weathering due to tracheophytes is (rightly) incorporated in the biogeochemical models which predict atmospheric CO₂ levels, so that these values of atmospheric CO₂ are not obtained independently of assumptions about plant behaviour in increasing the rate of weathering.

This largely vascular land plant-induced decrease in atmospheric CO₂ in the Devonian had important implications for performance of the aerial shoots in photosynthesis. Lower CO₂ partial pressures decrease the rates of C₃ photosynthesis over a range of photon flux densities if the anatomical and biochemical characteristics of the shoot are not modified and the near-extant O₂ partial pressures in the Devonian is considered (Beerling et al., 1998). Increased stomatal indices and/or densities with decreasing CO₂ partial pressures, as a result of adaptive or acclimatory processes, can partially offset the decreased CO₂ fixation rate, but at the expense of an increased rate of H₂O loss in transpiration per unit CO₂ fixed.

What evidence is there as to evolutionary changes inplant shoots in the Devonian in parallel with the tracheophyte (root)-induced drawdown of atmospheric CO₂? One change in shoot morphology through the Devonian which may be related to the decreased atmospheric CO₂ level (Chaloner, 1999) is the evolution of planate branch systems, and of ‘webbing’ of photosynthetic tissues between the ultimate ramuli of the branches (Gensel and Andrews, 1984; Stewart and Rothwell, 1993; Taylor and Taylor, 1993). Chaloner suggests that the decrease in mean diffusion pathlength for CO₂ in the gas phase of the intercellular gas spaces in the planate webbed structures than in the cylindrical photosynthetic structures can be related to the decreasing atmospheric CO₂ level through the Devonian (Chaloner, 1999). This could be related to the mechanical constraints on the dimensions of cylindrical as opposed to planar structures with a density of 10⁶ g m^-3 in at atmosphere with a density of 10³ g m^-3. At all events, the radius of cylindrical plant structures is greater than the smallest half-thickness of a laminar organ, a comparison which relates the maximum possible CO₂ diffusion pathlength from stomata to photosynthetic cells in the smallest diameter of cylindrical organ with that of the shortest CO₂ diffusion pathlength from stomata to photosynthetic cells in an amphistomatous planar structure. The decreased capacity for use of light energy in CO₂ fixation per unit projected area in a thin laminar structure relative to a small-diameter cylindrical organ can be countered at a whole-canopy level by increasing the number of layers (leaf area index: Raven, 1997, 1984a, b, 1993, 1994a, b, 1995a, 1997). Constraints on the structure of laminar organs in relation to the maximum possible distance between vascular tissue and the most distant photosynthetic cells have been discussed previously (Raven, 1994a). Clearly there are vascular plants today (when CO₂ levels in the atmosphere are similar to the lowest values seen in the Palaeozoic (Carboniferous)) which have cylindrical photosynthetic structures, frequently co-existing with tracheophytes with laminar photosynthetic structures so, as with so much in biology, there may be more than one structural type working well in performing a given function.

The other evolutionary change which occurred between the Silurian and the Carboniferous is an increase in stomatal density (Chaloner and Collinson, 1975; McElwain and Chaloner, 1995, 1996; Chaloner, 1999). The very wide genotypic range in stomatal densities among extant plants (Chaloner and Collinson, 1975), as judged from the wide stomatal density range for organisms all living in the same CO₂ environment means that extant comparators (NLE or nearest living equivalents) are needed in interpreting the fossil stomatal density values. This question has been addressed (McElwain, 1998; McElwain and Chaloner, 1995, 1996). This problem of NLEs occurs even for the recent past (Tertiary, and even the Mesozoic) when there were fossil plants with close phylogenetic similarity to extant plants (even extending to con-generics such as Ginkgo species from the recent and Mesozoic) as well as life-form and habitat similarities. For the Palaeozoic, close phylogenetic similarities do not occur, so the NLE comparators are largely those organisms with similar life forms and habitats.

With these provisos in mind (Edwards, 1998; McElwain, 1998; McElwain and Chaloner, 1995, 1996) a number of Palaeozoic plants have been analysed and a very significant decline in stomatal density between the Silurian and the Carboniferous shown. This is consistent with the decrease in CO₂ suggested by the Geocarb II model (Berner, 1994) and the theoretical and observed acclimatory effects of growth CO₂ on stomatal density in extant plants (Woodward, 1987; Woodward and Bazzaz, 1988; Woodward and Kelly, 1995; Beerling et al., 1998). The stomatal density changes over geological time have also been used to hindcast CO₂ levels (McElwain, 1998); this contrasts with the way they are used here.

These two examples, i.e. the evolution of laminar photosynthetic structures and of decreasing stomatal density in the Devonian, show the likely feedback of root activities, via atmospheric CO₂ drawdown, on shoot evolution.

	Conclusions

Top Abstract Introduction What is a root? The functions of roots... The biogeochemical context of... The early (Palaeozoic) evolution... Implications for plants and... Conclusions References

 
An understanding of the evolution of roots based on fossil evidence is only being obtained slowly. This fossil evidence does not support a monophyletic origin of roots in polysporangiophytes. Molecular genetic studies on root development in extant lycophytes and euphyllophytes (including seed plants) should give further information on this point.

	Acknowledgments

 
Work in JAR's laboratory on roots has been, and/or is, supported by AFRC, BBSRC and NERC, and DE's research on early land plants by NERC. Comments from Dr KR Skene and Dr W Bruce, and by Professor WG Chaloner and another, anonymous, referee have been very helpful.

	Notes

 
³ To whom correspondence should be addressed. Fax: +44 1382 344275. E-mail: j.a.raven{at}dundee.ac.uk

	References

Top Abstract Introduction What is a root? The functions of roots... The biogeochemical context of... The early (Palaeozoic) evolution... Implications for plants and... Conclusions References

 
Addicott FT. 1982. Abscission. Berkely: University of California Press.

Algeo TJ, Berner RA, Maynard JB, Sheckler SE. 1995. Late Devonian oceanic events and biotic crises: ‘rooted’ in the evolution of vascular land plants. GSA Today 45, 64–67.

Algeo TJ, Sheckler SE. 1998. Terrestrial-marine teleconnections in the Devonian: links between the evolution of land plants, weathering processes and marine anoxic events. Philosophical Transactions of the Royal Society of London B 353, 13–130.

Allaway WG, Ashford AE. 1996. Structure of hair roots of Lysinema ciliatum R.B. and its implications for their water relations. Annals of Botany 77, 383–388.

Aloni R, Enstone DE, Peterson C. 1998. Indirect evidence for bulk water flow in root cortical cells of three dicotyledonous species. Planta 207, 1–7.

Amos LA, Amos WB. 1991. Molecules of the cytoskeleton. Basingstoke and London: MacMillan Education Ltd.

Ashford AE, Allaway WG, Reed ML. 1996. A possible role for the thick-walled epidermal cells in the mycorrhizal hair roots of Lysinema ciliatum R. Br and other Epacridaceae. Annals of Botany 77, 375–381.

Atger C, Edelin CL. 1994. Preliminary data on the comparative architecture of roots and crowns. Canadian Journal of Botany 72, 963–975.

Atkinson D. 1992. How long is the life span of a root? Trends in Ecology and Evolution 7, 173–174.

Banks HP. 1970. Evolution and plants of the past. London: MacMillan.

Barlow PW. 1973. Mitotic cycles in root meristems. In: Balls M, Billett FS, eds. The cell cycle in development and differentiation. Cambridge University Press, 133–165.

Bateman RM, Crane PR, Di Michele WA, Kenrick PR, Rowe NP, Speck T, Stein WE. 1998. Early evolution of land plants: phylogeny, physiology and ecology of the primary terrestrial radiation. Annual Review of Ecology and Systematics 29, 263–192.

Baylis GTS. 1975. The magnolioid mycorrhiza and mycotrophy in root systems derived from it. In: Sanders FE, Mosse B, Tinker PB, eds. Endomycorrhizas. London: Academic Press, 373–389.

Beerling DJ, Woodward FI, Lomas MR, Wills HA, Quick WP, Valdes PJ. 1998. The influence of Carboniferous palaeatmospheres on plant function: an experimental and modelling assessment. Philosophical Transactions of the Royal Society of London B 353, 131–140.

Bell AD. 1991. Plant form. An illustrated guide to flowering plant morphology. Oxford University Press.

Benfey PN, Linstead PJ, Roberts K, Schiefelbein JW, Hauser M-T, Aeschbacher RA. 1993. Root development in Arabidopsis: four mutants with dramatically altered root morphogenesis. Development 119, 53–70.

Berner EK, Berner RA. 1996. Global environment. Water, air and geochemical cycles. New Jersey: Prentice Hall, Upper Saddle River.

Berner RA. 1990. Carbon dioxide over Phanerozoic time. Science 249, 1382–1386.

Berner RA. 1993. Palaeozoic atmospheric CO₂: importance of solar radiation and plant evolution. Science 261, 68–70.

Berner RA. 1994. Geocarb II: a revised model of atmospheric CO₂ over Phanerozoic time. American Journal of Science 289, 333–361.

Berner RA. 1997. The rise of plants and their effects on weathering and atmospheric CO₂. Science 276, 544–545.

Berner RA. 1998. The carbon cycle and CO₂ over Phanerozoic time: the role of land plants. Philosophical Transactions of the Royal Society of London B 353, 75–82.

Berntson GM. 1994. Modelling root architecture: are there tradeoffs between efficiency and potential of resource acquisition? New Phytologist 127, 483–493.

Bertrand I, Hinsinger P, Jaillard B, Arvieu JC. 1999. Dynamics of phosphorus in the rhizosphere and rape grown in synthetic, phosphated calcite and goethite. Plant and Soil 211, 111–119.

Bierhorst DW. 1971. Morphology of vascular plants. New York: The MacMillan Co, London: Collier-MacMillan.

Blackwell M. 2000. Terrestrial life—fungal from the start? Science 289, 1884–1885.

Boot RGA. 1989. The significance of size and morphology of root systems for nutrient acquisition and competition. In: Lambers H, Cambridge ML, Konings H, Pons TL, eds. Causes and consequences of variation in growth rate and productivity of higher plants. The Netherlands: SPB Academic Publishers bv, The Hague, 299–311.

Bret-Harte MS, Silk WK. 1994. Non-vascular, symplasmic diffusion of sucrose cannot satisfy the carbon demand of growth in the primary root tip of Zea mays. Plant Physiology 105, 19–33.

Cahn MD, Zobel RW, Bouldin DR. 1989. Relationship between root elongation rate and diameter and duration of growth of lateral roots of maize. Plant and Soil 119, 271–279.

Caldwell MM, Richards JH. 1986. Competing root systems: morphological models of absorption. In: Givnish TJ, ed. On the economy of plant form and function. Cambridge University Press, 251–273.

Canny MJ. 1973. Phloem translocation. Cambridge University Press.

Chaloner WG. 1970. The rise of the first land plants. Biological Reviews 45, 353–377.

Chaloner WG. 1999. Leaf structure and stomatal density as records of past CO₂ levels. Abstracts of XVI International Botanical Congress, St Louis, USA, 1–7 August 1999, p. 161.

Chaloner WG, Collinson ME. 1975. Application of SEM to a sigillarian impression fossil. Review of Palaeobotany and Palynology 20, 85–101.

Chaloner WG, Sheerin A. 1979. Devonian macrofloras. In: The Devonian system (Special Papers in Paelaeobotany 23). London: The Palaeontological Association, 145–161.

Clarkson DT. 1985. Factors affecting mineral nutrient acquisition by plants. Annual Review of Plant Physiology 36, 77–115.

Clarkson DT, Kuiper PJC, Lüttge U. 1986. Mineral nutrition: sources of nutrients for land plants from outside the pedosphere. Progress in Botany 48, 80–96.

Cutter EG. 1971. Plant anatomy. Experiment and interpretation. Part 2. Organs. London: Edward Arnold.

Damus M, Peterson RL, Enstone DE, Peterson CA. 1997. Modification of cortical cell walls in roots of seedless vascular plants. Botanica Acta 110, 190–195.

Darrah PR. 1993. The rhizosphere and plant nutrition: a quantitative approach. Plant and Soil 155/156, 1–20.

Dolan L, Scheres B. 1998. Root pattern: shooting in the dark? Cell and Developmental Biology 9, 201–206.

Driese SG, Mora CI. 2000. Diversification of Siluro-Devonian plant traces in palaeosols and influence on estimates of palaeoatmospheric CO₂ levels. In: Gensel PG, Edwards D, eds. Plants invade the land: evolutionary and environmental perspectives. New York: Columbia University Press, 237–253.

Duckett JG, Russell AJ, Ligrone R. 1996. Trichomes in the Hymenophyllaceae. In: Camus JM, Gibby M, John RJ, eds. Pteridology in perspective. Kew: Royal Botanic Gardens, 511–514.

Eames AJ. 1936. Morphology of vascular plants. Lower groups. New York: MacGraw-Hill.

Eames AJ, MacDaniels LH. 1947. An introduction to plant anatomy, 2nd edn. New York: MacGraw-Hill.

Edwards D. 1993. Cells and tissues in the vegetative sporophytes of early land plants. New Phytologist 125, 225–247.

Edwards D. 1994. Towards an understanding of pattern and process in the growth of early vascular plants. In: Ingram DS, Hudson A, eds. Shape and form in plants and fungi (Linnean Society Symposium Series No. 16). London: Academic Press, 35–59.

Edwards D. 1998. Climate signals in Palaeozoic land plants. Philosophical Transactions of the Royal Society of London B 353, 141–157.

Edwards D. 2000. The role of Mid-Palaeozoic mesofossils in the detection of early bryophytes. Philosophical Transactions of the Royal Society of London B 355, 733–755.

Edwards D, Abbott GD, Raven JA. 1996. Cuticles in early land plants: a palaeoecophysiological evaluation. In: Kierstens G, ed. Plant cuticles. Oxford: Bios Scientific Publishers, 1–31.

Edwards D, Duckett JG, Richardson JB. 1995. Hepatic characters in the earliest land plants. Nature 374, 635–636.

Edwards D, Wellman CH, Axe L. 1998. The fossil record of early plants and interrelationships between primitive embryophytes: too little and too late? In: Bates JW, Ashton NW, Duckett JG, eds. Bryology for the twenty-first century. Leeds: British Bryological Society/Maney Press, 15–43.

Eissenstadt DM, Yanai RD. 1997. The ecology of root lifespan. In: Fitter AM, ed. Advances in ecological research, Vol. 27. San Diego: Academic Press, 1–60.

Elick JM, Driese SG, Mora CI. 1998. Very large plant and root traces from the Early and Middle Devonian: implications for early terrestrial ecosystems and atmospheric p(CO₂). Geology 26, 143–146.

Ennos AR. 1993. The function and formation of buttresses. Trends in Ecology and Evolution 8, 350–351.

Ennos AR, Fitter AH. 1992. Comparative functional morphology of the anchorage systems of annual dicots. Functional Ecology 6, 71–78.

Epstein E. 1999. Silicon. Annual Review of Plant Physiology and Plant Molecular Biology 50, 641–664.

Esau K. 1953. Plant anatomy. New York: John Wiley & Sons Inc.

Fahn A. 1974. Plant anatomy, 2nd edn. Oxford: Pergamon Press.

Fairon-Demart M, Li C-S. 1993. Lorophyton goense gen. et sp. nov. from the Lower Givetian of Belgium and a discussion of Lower Devonian Cladoxylopsida. Review of Palaeobiology and Palynology 77, 1–22.

Fitter AH. 1987. An architectural approach to the comparative ecology of plant root systems. In: Rorison IH, Grime JP, Hunt R, Hendry GAF, Lewis DH, eds. Frontiers of comparative plant ecology. London: Academic Press, 217–233.

Fitter AH. 1996. Characteristics and functions of roots systems. In: Waisel Y, Eshel A, Kafkafi U, eds. Plant roots: The hidden half, 2nd edn. New York: Marcel Dekker, 1–20.

Fitter AH, Moyersoen B. 1996. Evolutionary trends in root-microbe symbioses. Philosophical Transactions of the Royal Society of London B 351, 1367–1375.

Gensel PG, Andrews HN. 1984. Plant life in the Devonian. New York: Praeger.

Gensel PG, Kotyk M, Basinger JF. 2001. Morphology of above- and below-ground structures in early Devonian (Pragian-Emsian). In: Gensel PG, Edwards D, eds. Plants invade the land: evolutionary and environmental perspectives. New York: Columbia University Press, 83–102.

Gifford EM, Foster AS. 1987. Morphology and evolution of vascular plants, 3rd edn. New York: WH Freeman and Company.

Gill AM, Tomlinson PB. 1975. Aerial roots. An array of forms and functions. In: Torrey JG, Clarkson DT, eds. The development and function of roots. London: Academic Press, 237–260.

Gilroy S, Jones DL. 2000. Through form to function: root hair development and nutrient uptake. Trends in Plant Science 5, 56–60.

Goebel K. 1930. Organographie der Pflanzen, 2nd edn, Part 2. Jena: G. Fischer.

Golubic S, Seong-Joo L. 1999. Early cyanobacterial fossil record: preservation, palaeoenvironments and identification. European Journal of Phycology 34, 339–348.

Graham LE. 1993. Origin of land plants. New York: John Wiley & Sons.

Gray J. 1985. The microfossil record of early land plants: advances in understanding of early terrestrialization, 1970–1984. In: Chaloner WG, Lawson JD, eds. Evolution and environment in the late Silurian and Early Devonian. Philosophical Transactions of the Royal Society of London B 309, 167–195.

Groff PA, Kaplan DR. 1988. The relation of root systems to shoot systems in vascular plants. Botanical Reviews 54, 387–422.

Grubb PJ. 1961. Uptake and movement of salts in Polytrichum. PhD thesis, University of Cambridge.

Grubb PJ. 1970. Observations on the structure and biology of Haplomitrium and Takakia, hepatics with roots. New Phytologist 69, 303–326.

Gualtieri G, Bisseling T. 2000. The evolution of nodulation. Plant Molecular Biology 42, 181–194.

Harper JL. 1977. Population biology of plants. London: Academic Press.

Harrison MJ, van Buuven ML. 1995. A phosphate transporter from the mycorrhizal fungus Glomus versiforme. Nature 378, 626–629.

Hawes MC, Gunawardena U, Miyasaka S, Zhao X. 2000. The role of root border cells in plant defense. Trends in Plant Science 5, 128–133.

Hawksworth DL. 1988. The variety of fungal–algal symbioses, their evolutionary significance and the nature of lichens. Botanical Journal of the Linnean Society 96, 3–20.

Herwitz SR. 1991. Above-ground adventitious roots and stemflow chemistry of Ceratopetalum virihowii in an Australian montane tropical rain-forest. Biotropica 23, 210–218.

Hodge A, Robinson D, Fitter DM. 2000. An arbuscular mycorrhizal inoculum enhances root proliferation but not nitrogen capture from nutrient-rich patches in soil. New Phytologist 145, 575–584.

Horodyski RJ, Knauth LP. 1994. Life on land in the Precambrian. Science 263, 494–498.

Hotton CL, Hueber FM, Griffing DH, Bridge JS. 2001. Early terrestrial plant environments: an example from the Emsian of Gaspé, Canada. In: Gensel PG, Edwards D, eds. Plants invade the land: evolutionary and environmental perspectives. New York: Columbia University Press, 179–212.

Hueber FM. 1983. A new species of Baragwanathia from the Sextant Formation (Emsian), Northern Ontario, Canada. Botanical Journal of the Linnean Society 86, 57–79.

Jackson RB, Mooney HA, Schulze ED. 1997. A global budget for fine root biomass, surface area and nutrient contents. Proceedings of the National Academy of Sciences, USA 94, 7362–7366.

Jones DL, Darrah PR. 1993. Influx and efflux of amino acids from Zea mays L. roots and their implications for N nutrition and the rhizosphere. Plant and Soil 155/157, 87–90.

Jones DL, Edwards, AC, Donachie K, Darrah PR. 1994. Role of proteinaceous amino acids released in root exudates in nutrient acquisition from the rhizosphere. Plant and Soil 158, 183–192.

Julian D, Gaill F, Wood E, Arp AJ, Fisher CR. 1999. Roots as a site of hydrogen sulfide uptake in the hydrocarbon seep vestimentiferan Lamellibrachia spp. Journal of Experimental Biology 202, 2245–2257.

Keeley JE, Osmond CB, Raven JA. 1984. Stylites, a vascular land plant without stomata absorbs CO₂ via its roots. Nature 310, 694–695.

Keller CK, Wood BD. 1993. Possibility of chemical weathering before the advent of vascular plants. Nature 464, 407–412.

Kenrick P, Crane PR. 1997. The origin and early diversification of land plants. A cladistic analysis. Washington: Smithsonian Institution Press.

Kerp H, Hass H, Mossbragger V. 2001. New data on Nothia aphylla Lyon 1964 ex El-Saadawy et Lacey 1997, a poorly known plant from the Lower Devonian Rhynie Chert, In: Gensel PG, Edwards D, eds. Plants invade the land: evolutionary and environmental perspectives. New York: Columbia University Press, 52–82.

Kirk GJD, Santos EE, Findenegg GR. 1994. Phosphate solubilization by organic anion excretion from rice (Oryza sativa L.) growing in aerobic soil. Plant and Soil 211, 11–18.

Kochian LV. 1995. Cellular mechanisms of aluminium toxicity and resistance in plants. Annual Review of Plant Physiology and Plant Molecular Biology 46, 237–260.

Konrad W, Roth-Nebesick A, Kerp H, Hass H. 2000. Transpiration and assimilation of Early Devonian land plants with axially symmetric telomer—simulations on the tissue level. Journal of Theoretical Biology 206, 91–107.

Li CS, Edwards D. 1995. A reinvestigation of Halle Drepanophycus spinaeformis Gopp from Lower Devonian of Yunnan Province, Southern China. Botanical Journal of the Linnean Society 118, 163–192.

Lovelock JE, Maggs RJ, Rasmusen RA. 1972. Atmospheric dimethylsulphide and the natural sulphur cycle. Nature 237, 452–453.

Lu PZ, Jernstedt JA. 1996. Rhizophore and root development in Selaginella martensii. Meristem variations and identity. International Journal of Plant Science 157, 180–194.

Lyndon RF. 1973. The cell cycle in the shoot apex. In: Balls M, Billett FS, eds. The cell cycle in development and differentiation. Cambridge University Press, 167–183.

Marschner H. 1995. Mineral nutrition of higher plants, 2nd edn. London: Academic Press.

Mathesius V, Schlaman HRM, Spaink HP, Sautter C, Rolfe BG, Djordjevec MA. 1998. Auxin transport inhibition precedes root nodule formation in white clover roots and is regulated by flavonoids and derivatives of chitin oligosaccharides. The Plant Journal 14, 23–34.

McCully ME. 1999. Roots in soil: unearthing the complexities of roots and their rhizospheres. Annual Review of Plant Physiology and Plant Molecular Biology 50, 695–718.

McElwain JC. 1998. Do fossil plants signal palaeoatmosphere CO₂ concentration in the geological past? Philosophical Transactions of the Royal Society of London B 353, 83–96.

McElwain JC, Chaloner WG. 1995. Stomatal density and index of fossil plants track atmospheric carbon dioxide in the Palaeozoic. Annals of Botany 76, 389–395.

McElwain JC, Chaloner WG. 1996. The fossil cuticle as a skeletal record of environmental change. Palaios 11, 376–388.

Meyer-Berthaud B, Sheckler SE, Wendt J. 1999. Archaeopteris is the earliest known modern tree. Nature 398, 700–701.

Moteetee A, Duckett JG, Russell AJ. 1996. Mycorrhizas in the ferns of Letsotho. In: Camus JM, Gibby M, Johns RJ, eds. Pteridology in perspective. Kew: Royal Botanic Gardens, 621–631.

Neumann G, Römheld V. 1999. Root excretion of carboxylic acids and protons in phosphorus-deficient plants. Plant and Soil 211, 121–130.

Nicolson TH. 1975. Evolution of vesicular-arbuscular mycorrhizas. In: Saunders FE, Mosse B, Sanders PB, eds. Endomycorrhizas. London: Academic Press, 25–34.

Niklas KJ. 1992. Plant biomechanics. An engineering approach to plant form and function. Chicago: University of Chicago Press.

Niklas KJ. 1994. Plant allometry. The scaling of form and process. Chicago and London: The University of Chicago Press.

Niklas KJ. 1997. The evolutionary biology of plants. Chicago and London: The University of Chicago Press.

Page CN. 1982. The ferns of Britain and Ireland. Cambridge University Press.

Parsons R, Raven JA, Sprent JI. 1995. Translocation of iron to the N₂-fixing stem nodules of Sesbania rostrata Brem. Journal of Experimental Botany 46, 291–296.

Parsons R, Sprent JI, Raven JA. 1993. Humidity and light affect the growth, development and nitrogenase activity of stem nodules of Sesbania rostrata Brem. New Phytologist 125, 749–755.

Passioura JB, Ashford AE. 1974. Rapid translocation in the phloem of wheat roots. Australian Journal of Plant Physiology 1, 521–527.

Peterson CA, Cholewa E. 1998. Structural modifications of the apoplast and their potential impact on ion uptake. Zeitschrift für Pflanzenernahrung und Bodenkund 161, 521–531.

Peterson RL. 1992. Adaptation of root structure in relation to biotic and abiotic factors. Canadian Journal of Botany 70, 661–675.

Pirozynski KA, Malloch DW. 1975. The origin of land plants: a matter of mycotrophism. Biosystems 6, 153–164.

Pirozynski KA, Dalpé Y. 1989. Geological history of the Glomaceae with particular reference to mycorrhizal symbiosis. Symbiosis 7, 1–36.

Plassard C, Melsem M, Souche G, Jaillard B. 1999. Localization and quantification of net fluxes of H⁺ along maize roots by combined use of pH-indicator dye videodensitometry and H⁺-selective microelectrodes. Plant and Soil 211, 29–39.

Pritchard J. 1994. The control of cell expansion in roots. New Phytologist 127, 3–26.

Qui YL, Lee J, Bernasconi-Quadroni F, Soltis DE, Soltis PS, Zanis M, Zimmer EA, Chen Z, Savoliner V, Chase HW. 1999. The earliest angiosperms: evidence from mitochondrial, plastid and nuclear genes. Nature 402, 404–407.

Raven JA. 1970. Exogenous inorganic carbon sources in plant photosynthesis. Biology Reviews 45, 167–221.

Raven JA. 1977. The evolution of vascular land plants in relation to supracellular transport processes. In: Woolhouse HW, ed. Advances in botanical research, Vol. 5. London: Academic Press, 153–219.

Raven JA. 1981. Nutritional strategies of submerged benthic plants: the acquisition of C, N and P by rhizophytes and haptophytes. New Phytologist 92, 1–20.

Raven JA. 1983. The transport and function of silicon in plants. Biological Reviews 58, 179–207.

Raven JA. 1984a. The comparative physiology of the adaptation of plants and arthropods to life on land. Philosophical Transactions of the Royal Society of London B 309, 273–288.

Raven JA. 1984b. Physiological correlates of the morphology of early vascular plants. Botanical Journal of the Linnean Society 88, 105–126.

Raven JA. 1984c. Energetics and transport in aquatic plants. New York: AR Liss.

Raven JA. 1986. The evolution of plant life forms. In: Givnish TJ, ed. On the ecology of plant form and function. Cambridge University Press, 421–492.

Raven JA. 1989. Overview of transport systems in algae and bryophytes. In: Fleischer S, Fleischer B, eds. Methods in enzymology: biomembranes/biological transport, Vol. 4, Part V (Vol. 174). San Diego: Academic Press, 366–390.

Raven JA. 1991. Terrestrial rhizophytes and H⁺ currents circulating over at least a millimetre: an obligate relationship? New Phytologist 117, 177–185.

Raven JA. 1993. The evolution of vascular land plants in relation to quantitative functioning of dead water-conducting cells and stomata. Biological Reviews 68, 337–363.

Raven JA. 1994a. The significance of the distance from photosynthesising cells to vascular tissue in extant and early vascular plants. Botanical Journal of Scotland 47, 65–82.

Raven JA. 1994b. Physiological analyses of aspects of the functioning of vascular tissue in early plants. Botanical Journal of Scotland 47, 49–64.

Raven JA. 1995a. The early evolution of land plants: aquatic ancestors and atmospheric interactions. Botanical Journal of Scotland 47, 151–175.

Raven JA. 1995b. Symplasmic proton fluxes in photosynthesising and developing plant tissues. Biological Reviews 70, 189–224.

Raven JA. 1997. The role of marine biota in the evolution of terrestrial biota: gases and genes. Biogeochemistry 39, 139–164.

Raven JA. 1998. Extrapolating feedback processes from the present to the past. Philosophical Transactions of the Royal Society of London B 353, 19–28.

Raven JA. 1999a. The size of cells and organisms in relation to the evolution of embryophytes. Plant Biology 1, 2–12.

Raven JA. 1999b. The minimum size of seeds and spores in relation to the ontogeny of homoiohydry. Functional Ecology 13, 5–14.

Raven JA. 2000a. Land plant biochemistry. Philosophical Transactions of the Royal Society of London B 355, 833–846.

Raven JA. 2000b. Silicon transport at the cell and tissue level. In: Datnoff LE, Korndorfer GH, Syder GH, eds. Silicon in agriculture. Amsterdam: Elsevier (in press).

Raven JA, Yin Z-H. 1998. The past, present and future of nitrogenous compounds in the atmosphere and their interactions with plants. New Phytologist 139, 205–219.

Raven JA, Smith FA, Smith FE. 1980. Ions and osmoregulation. In: Rains DW, Valentine RC, Hollaender A, eds. Genetic engineering of osmoregulation: impact of plant productivity for food, chemicals and energy. New York: Plenum Press, 101–108.

Raven JA, Franco AA, de Jesus EL, Jacob Neto J. 1990. H⁺ extrusion and organic acid synthesis in N₂-fixing symbioses involving vascular plants. New Phytologist 114, 369–389.

Raven JA, Johnston AM, Kübler J, Parsons R. 1994. The influence of natural and experimental high O₂ concentrations on O₂-evolving photolithotrophs. Biological Reviews 69, 61–94.

Raven JA, Griffiths H, Smith EC, Vaughn KC. 1998. New perspectives in the biophysics and physiology of bryophytes. In: Bates JW, Ashton NW, Duckett JG, eds. Bryology for the twenty-first century. Leeds: British Bryological Society/Maney Publishing, 261–275.

Raven JA, Handley LL, MacFarlane JJ, McInroy S, McKenzie L, Richards JH, Samuelsson G. 1988. The role of root CO₂ uptake and CAM in inorganic C acquisition by plants of the isoetid life form. A review, with new data on Ericocaulon decangulare. New Phytologist 108, 125–148.

Rayner RJ. 1983. New observations on Sawdonia ornata from Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences 74, 79–93.

Rayner RJ. 1984. New finds of Drepanophycus spinaeformis Göppert from the Lower Devonian of Scotland. Transactions of the Royal Society of Edinburgh: Earth Sciences 75, 353–363.

Read DJ. 1999. The ecophysiology of mycorrhizal symbioses with special reference to impacts upon plant fitness. In: Press MC, Scholes JD, Barker MG, eds. Physiological plant ecology. Symposium Number 39 of the British Ecological Society. Oxford: Blackwell, 133–152.

Read DJ, Duckett JG, Francis R, Ligrone R, Russell A. 2000. Symbiotic fungal associations in ‘lower’ land plants. Philosophical Transactions of the Royal Society of London B 355, 815–831.

Redecker D, Kodner R, Graham LE. 2000. Glomalean fungi from the Ordovician. Science 289, 1920–1921.

Remy W, Remy D, Hass H. 1997. Organisation, Wuchsformen und Lebensstrategien früher Landpflanzen des Unterdevens. Botanisches Jahrbuch Systematik 119, 509–562.

Retallack GJ. 1997. Early forest soils and their role in Devonian global change. Science 276, 583–585.

Richardson K, Griffiths H, Reed ML, Raven JA, Griffiths NA. 1984. Inorganic C assimilation in the Isoetids, Isoetes lacustris L. and Lobelia dortmanna L. Oecologia 61, 115–121.

Robinson D. 1990. Phosphorus availability and cortical senescences in cereal roots. Journal of Theoretical Biology 145, 257–265.

Robinson D. 1991. Roots and resource fluxes in plants and communities. In: Atkinson D, ed. Plant root growth. An ecological perspective. Special Publication Number 10 of the British Ecological Society. Oxford: Blackwell, 103–130.

Robinson D. 1996. Resource capture by localized root proliferation: why do plants bother? Annals of Botany 77, 179–185.

Robinson D, Linehan DJ, Caul S. 1991. What limits nitrate uptake from soil? Plant, Cell and Environment 14, 77–85.

Robinson D, Hodge A, Griffiths BS, Fitter AH. 1998. Plant root proliferation in nitrogen-rich patches confer competitive advantage. Proceedings of the Royal Society of London B 266, 431–435.

Rothwell GW. 1984. The apex of Stigmaria (Lycopsida), rooting organ of Lepidodendrales. American Journal of Botany 71, 1031–1034.

Rothwell GW. 1996. Phylogenetic relationships of ferns: a palaeobotanical perspective. In: Camus JM, Gibby M, John RJ, eds. Pteridology in perspective. Kew: Royal Botanic Gardens, 395–404.

Rothwell GW, Erwin DM. 1985. The rhizomorph apex of Paurodendron, implications for homologies among the rooting organs of Lycopsida. American Journal of Botany 72, 86–98.

Rye R, Holland HD. 2000. Life associated with a 2.76 Ga ephemeral pond? Evidence from Mount Roe #2 palaeosol. Geology 28, 483–486.

Sand-Jensen K, Prahl C, Stokholm H. 1982. Oxygen release from roots of submerged aquatic macrophytes.Oikos 38, 349–354.

Scheres B, Benfey PN. 1999. Asymmetric cell division in plants. Annual Review of Plant Physiology and Plant Molecular Biology 50, 505–537.

Schlesinger WH. 1991. Biogeochemistry: an analysis of global change. New York: Academic Press.

Schneider H. 1996. The root anatomy of ferns: a comparative study. In: Camus JM, Gibby M, John RJ, eds. Pteridology in perspective. Kew: Royal Botanic Gardens, 271–283.

Schweitzer H-J. 1980. Über Drepanophycus spinaeformis Göppert. Bonner Paläobotanische Mitteilung 7, 1–29.

Scott R. 1977. Plant root systems. Their function and interaction with the soil. London: McGraw Hill Book Co.

Shear WA. 1991. The early development of terrestrial ecosystems. Nature 351, 283–289.

Sherwood-Pike MA, Gray J. 1985. Silurian fungal remains: probable records of the class Ascomycetes. Lethaia 18, 1–20.

Sievers A, Schnepf E. 1981. Morphogenesis and polarity of tubular cells with tip growth. In: Kiermeyer O, ed. Cytomorphogenesis in plants. Wien: Springer Verlag, 265–299.

Silberbush M, Barber SA. 1983. Sensitivity of similated phosphorus uptake to parameters used by a mechanistic mathematical model. Plant and Soil 74, 93–100.

Simon L. 1996. Phylogeny of the Glomales: deciphering the past to understand the present. New Phytologist 133, 95–101.

Simon L, Bousquet J, Lévesque RC, Lalande M. 1993. Origin and diversification of endomycorrhizal fungi are coincidental with vascular land plants. Nature 363, 67–69.

Skene KR. 1998. Cluster roots: some ecological considerations. Journal of Ecology 86, 1060–1064.

Skene KR. 2000. Pattern formation in cluster roots: some developmental and evolutionary considerations. Annals of Botany 85, 901–908.

Smith DK, Davison PG. 1993. Antheridia and sporophytes in Takakia ceratophylla (Mitt.) Grelle: evidence for reclassification among the mosses. Journal of the Hattori Botanical Laboratory 73, 263–271.

Smith GM. 1938. Cryptogamic botany, Vol. II. Bryophytes and Pteridophytes. New York: McGraw-Hill.

Smith SE, Read DH. 1997. Mycorrhizal symbiosis, 2nd edn. San Diego: Academic Press.

Soltis PS, Soltis DE, Chase MW. 1999. Angiosperm phylogeny inferred from multiple genes as a tool for comparative biology. Nature 402, 402–404.

Spicer RA. 1989. Physiological characteristics of land plants in relation to environment through time. Transactions of the Royal Society of Edinburgh, Earth Sciences 80, 321–329.

Sprent JI. 1989. Which steps are essential for the formation of functional legume nodules? New Phytologist 111, 129–153.

St John TV. 1980. Root size, root hairs and mycorrhizal infection: a re-examination of Baylis' hypothesis with tropical trees. New Phytologist 84, 483–487.

Stebbins GC, Hill GJC. 1980. Did multicellular plants invade the land? The American Naturalist 115, 342–353.

Stebbins H, Hyams JS. 1979. Cell motility. London: Longman.

Stewart WN, Rothwell GW. 1993. Palaeobotany and the evolution of plant, 2nd edn. Cambridge University Press.

Strother PK. 2000. The cryptospore record indicates a Cambrian origin for land plants. American Journal of Botany 87, 77s–78s.

Taylor TN. 1994. The fossil history of Ascomycetes. In: Hawksworth DL, ed. Ascomycete systematics: Problems and perspectives in the nineties. New York: Plenum Press, 167–174.

Taylor TN, Taylor EI. 1993. The biology and evolution of fossil plants. New York: Prentice Hall.

Taylor TN, Remy W, Hass H, Kerp H. 1995. Fossil arbuscular mycorrhizae from the Early Devonian. Mycologia 87, 560–573.

Taylor TN, Hass H, Kerp H. 1997. A cyanolichen from the Lower Devonian Rhynie Chert. American Journal of Botany 84, 992–1004.

Thumfort PP, Atkins CA, Layzell DB. 1994. A re-evaluation of the role of the infected cell in the control of O₂ diffusion in legume nodules. Plant Physiology 105, 1329–1333.

Tidwell WD. 1998. Common fossil plants of Western North America, 2nd edn. New York: Smithsonian Institution Press.

Varney GT, Canny MJ. 1993. Rates of water uptake into the mature root system of maize plants. New Phytologist 123, 775–789.

Varney GT, Canny MJ, Wang XL, McCully ME. 1991. The branch roots of Zea. I. First order branches, their number, sizes and division into classes. Annals of Botany 67, 357–364.

Varney GT, McCully ME. 1991. The branch roots of Zea. II. Developmental loss of the apical meristem in field-grown roots. New Phytologist 118, 535–546.

Vestal JR. 1993. Cryptoendolithic communities from hot and cold deserts: speculation on microbial colonization and succession. In: Miles J, Walter DHO, eds. Primary succession on land. Special Publication Number 12 of the British Ecological Society. Oxford: Blackwell Scientific Publications, 5–16.

Walton DWH. 1993. The effects of cryptograms on mineral substrates. In: Miles J, Walter DHO, eds. Primary succession on land. Special Publication Number 12 of the British Ecological Society. Oxford: Blackwell Scientific Publications, 33–53.

Walton J. 1940. An introduction to the study of fossil plants. London: A C Black.

Wellman GH, Gray J. 2000. The microfossil record of early land plants. Philosophical Transactions of the Royal Society of London B 355, 717–732.

White PJ, Bowen KE, Woolaway KE. 2000. Transport of calcium to the xylem in plant roots. Journal of Experimental Botany 51, Supplement, 5.

von Wirén N, Lauter F-R, Ninemann O, Gillissen B, Walch-Liu P, Engels C, Jost W, Frommer WB. 2000. Differential regulation of three functional ammonium transporter genes by nitrogen in root hairs and by light in leaves of tomato. The Plant Journal 21, 167–175.

Woodward FI. 1987. Stomatal numbers are sensitive to increases in CO₂ from pre-industrial levels. Nature 327, 617–618.

Woodward FI, Bazzaz FA. 1988. The response of stomatal density to CO₂ partial pressure. Journal of Experimental Botany 39, 1771–1781.

Woodward FI, Kelly CK. 1995. The influence of CO₂ concentration on stomatal density. New Phytologist 131, 311–327.

Wynn-Williams DD. 1993. Microbial processes and initial stabilization of fellfield soil. In: Miles J, Walter DHO, eds. Primary succession on land. Special Publication Number 12 of the British Ecological Society. Oxford: Blackwell Scientific Publications, 17–32.

Yapp CJ, Poths H. 1992. Ancient atmospheric CO₂ pressures inferred from natural goethites. Nature 355, 342–344.

Yapp CJ, Poths H. 1994. Productivity of pre-vascular continental biota inferred from the Fe(CO₃)OH contents of goethite. Nature 368, 49–51.

Zhang H, Forde BG. 1998. An Arabidopsis MADS-box gene that controls nutrient-induced changes in root architecture. Science 279, 407–409.

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				C. K. Boyce How green was Cooksonia? The importance of size in understanding the early evolution of physiology in the vascular plant lineage Paleobiology, March 1, 2008; 34(2): 179 - 194. [Abstract] [Full Text] [PDF]

				A. Pierret, C. Doussan, Y. Capowiez, F. Bastardie, and L. Pages Root Functional Architecture: A Framework for Modeling the Interplay between Roots and Soil Vadose Zone J., May 17, 2007; 6(2): 269 - 281. [Abstract] [Full Text] [PDF]

				M. Kennedy, M. Droser, L. M. Mayer, D. Pevear, and D. Mrofka Late Precambrian Oxygenation; Inception of the Clay Mineral Factory Science, March 10, 2006; 311(5766): 1446 - 1449. [Abstract] [Full Text] [PDF]

				D. J. BEERLING Leaf Evolution: Gases, Genes and Geochemistry Ann. Bot., September 1, 2005; 96(3): 345 - 352. [Abstract] [Full Text] [PDF]

				D. J. Beerling and R. A. Berner Feedbacks and the coevolution of plants and atmospheric CO2 PNAS, February 1, 2005; 102(5): 1302 - 1305. [Abstract] [Full Text] [PDF]

				W. E. Friedman, R. C. Moore, and M. D. Purugganan The evolution of plant development Am. J. Botany, October 1, 2004; 91(10): 1726 - 1741. [Abstract] [Full Text] [PDF]

				G. Hou, J. P. Hill, and E. B. Blancaflor Developmental anatomy and auxin response of lateral root formation in Ceratopteris richardii J. Exp. Bot., March 1, 2004; 55(397): 685 - 693. [Abstract] [Full Text] [PDF]

				J. Z. Kiss, K. M. Miller, L. A. Ogden, and K. K. Roth Phototropism and Gravitropism in Lateral Roots of Arabidopsis Plant Cell Physiol., January 1, 2002; 43(1): 35 - 43. [Abstract] [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) E-letters: Submit a response Alert me when this article is cited Alert me when E-letters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Disclaimer Google Scholar Articles by Raven, J. A. Articles by Edwards, D. Search for Related Content PubMed PubMed Citation Articles by Raven, J. A. Articles by Edwards, D. Agricola Articles by Raven, J. A. Articles by Edwards, D. Social Bookmarking          What's this?

Online ISSN 1460-2431 - Print ISSN 0022-0957

Copyright © 2005 Society for Experimental Biology

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics