The minimum size of seeds and spores in relation to the ontogeny of homoiohydric plants

Abstract

Homoiohydry may be defined as the capacity of non-dormant terrestrial plants to remain hydrated, albeit without the capacity for net CO2 fixation, when the soil can supply water to the plant less rapidly than the atmosphere can potentially remove water in transpiration. Homoiohydry is an attribute of the sporophytes of most extant vascular plants and involves significant differentiation with roots (or functionally analogous structures), xylem, cuticle, stomata and intercellular gas spaces (Raven 1977; Edwards, Abbott & Raven 1996; Bell & Mooers 1997). Homoiohydry was a major physiological innovation in the evolution of embryophytes and has enabled plants 1–100 m high to grow in terrestrial habitats with fluctuating soil water availability and evaporative demand of the atmosphere (Raven 1977, 1984a, 1993, 1997a, 1998). The possession of homoiohydric characteristics permits non-dormant sporophytic structures to be desiccation-intolerant, i.e. unable to survive equilibration with an atmospheric relative humidity of less than c. 50% (i.e. a water potential of c.–100 MPa). Whatever the mechanistic basis for this height restriction of desiccation-tolerant growing vegetative structures of land plants to less than 1 m, it appears that the occurrence of terrestrial plants in the 1–100 m height range requires desiccation-intolerance in the growing sporophyte combined with homoiohydry. It must be remembered that desiccation-tolerance (as defined above) is common in dormant structures (e.g. seeds) of those plants which are desiccation-intolerant when growing and apparently relates to the widespread occurrence in vascular plants of a number of closely related genes encoding ‘late-embryogenesis-abundant’ proteins, one group of which is termed the dehydrins (Ingrams & Bartels 1996). Major reproductive innovations in the evolution of embryophytes were the polyphyletic evolution of heterospory and, in a few of these parallel heterosporous lines of evolution, the seed habit, and culminated in the plant supplying all of the water needed to its reproductive structures, so that above-ground sexual reproductive processes relied only on soil water supply to the sporophyte rather than on rain or a dew making the plant (sporophyte or gametophyte) surface wet. These two (partial) emancipations of large terrestrial plants from a continuous supply of external water for vegetative and reproductive processes still leaves a life-cycle phase, that of seed germination, in which a continuous supply of water at or near the soil surface is necessary. How rapidly a seedling can establish itself as a homoiohydrically functional photosynthetic entity determines how long it has to spend with a net C and energy loss (respiration exceeding photosynthesis) and, to the extent that it has photosynthetic parts exposed to the atmosphere, water loss to the atmosphere without the benefit of the complete homoiohydric system which regulates hydration. The duration of this phase of vulnerability between the dormant, and frequently desiccation-tolerant, seed and the homoiohydrically competent, photosynthetically self-sufficient, seedling depends on two factors in addition to the rate at which seed reserves can be used to build a seedling. One factor is the minimum size at which the seedling is homoiohydrically competent and capable of net C and energy gain over a light–dark cycle, and the other is the size of the C and energy reserves in the seed for that majority of seeds which do not have ready access to external organic C. Sources of external organic C include parasitism of the seedling on its mycorrhizal fungi or on a photosynthetic higher plant. Another source of external organic C is the parent plant in the case of viviparous seedlings, e.g. those of the seagrasses Amphibolis spp. and Thalassodendron spp. (Kuo & Kirkman 1996) and, more relevant to terrestrial homoiohydric plants, in mangroves such as Rhizophora (Clough 1982; Tomlinson 1986) where the parent may supply not only organic C but also water used in transpiration and growth of the photosynthesizing viviparous seedling. In these cases any relation of seed size to the size of a homoiohydrically functional seedling is decoupled by external organic C inputs. Furthermore, in some of these cases the seedling and, indeed, the adult plant is not a homoiohydrically competent photosynthetic organism (the submerged seagrasses Amphibolis and Thalassodendron) or, for holoparasites such as Orobanche (Raven 1983) and ‘saprophytes’ (obligately parasitic on their mycorrhizal fungi, e.g. Monotropa: Smith & Read 1997), not even photosynthetic. The minimum size of seeds of seed plants with different nutritional modes for their seedlings is relatively well established observationally, with a dry mass of about 5 μg (Sonerila heterostemon Naud.) for seeds producing seedlings whose first non-seed energy source is photosynthesis (Harper, Lovell & Moore 1970; Grime et al. 1981; Westoby, Jurado & Leishman 1992; Metcalfe & Grubb 1995; Swanborough & Westoby 1996; Grubb et al. 1998; noting that the 1 μg for Sagina apetala Ard. cited by Grime et al. 1981 should be 10 μg: K. Thompson, personal communication to P. J. Grubb). Seeds as small as 1 μg dry mass are known when the seedlings they produce use some other living organism as their earliest (and sometimes only) non-seed energy source, i.e. those which live by mycotrophy (parasitism on a fungus which in turn is living parasitically or mutualistically on a photosynthetic plant, or saprotrophically) or by parasitism of a photosynthetic plant (Harper et al. 1970; Janzen 1983). No measurements of the minimum size of a photosynthetically self-sufficient, homoiohydrically competent seedling seem to have been made, nor do there seem to be any computations as to what such a minimum seedling size might be. This paper sets out to consider this question of the minimum size of such photosynthetically self-sufficient, homoiohydrically competent seedling by considering first the possible determinants of the minimum size of a photosynthetically self-sufficient, homoiohydrically competent seedling. This suggested minimum size for such a seedling is then compared with the minimum observed size of seeds yielding non-mycotrophic, non-parasitic seedlings as one possible determinant of the minimum size of such seeds. Finally, we deal with the phylogeny of homoiohydry and the seed habit to see if the deliberations on the minimum size of seeds and of homoiohydric seedlings can shed any light on the ecophysiological performance of young sporophytes of homosporous and heterosporous pteridophytes. Embryos of flowering plants typically only become green and hence photosynthetically active when exposed to light after germination. None of the very small seeds under consideration seem to depart from this paradigm by having ‘chloroembryos’ in mature seeds (with little or no dormancy) which photosynthesize in situ thanks to transparency of the fruit wall and seed tissues (Janzen 1982; Antoszewski, Dabydeen & Lalla 1989; cf. Asokanthan et al. 1997), or are able to synthesize chlorophyll in the dark (Adamson, Packer & Gregory 1988). Among the factors which determine the minimum size of homoiohydrically functional seedlings of a seed plant are the following. One is the occurrence of a certain minimum height of the plant spanning the soil–atmosphere interface, so that homoiohydry can occur with acquisition of water from the soil and its (regulated) loss to the atmosphere. A second factor which may impose a lower limit on seedling size concerns the functioning of the xylem. Even with a short distance of longitudinal transport of water through the ectohydric system there is a minimum radius (0·5–1·0 μm) of xylem element below which driving forces for water movement become almost as high as those needed to transport water through the surrounding parenchyma tissue (see Raven 1977, 1984a, 1993). Even with a single file of xylem elements (a rather unsafe procedure granted the possibility of embolism, cavitation or collapse of elements during the transport of solutions under tension, but one that is borne by some successful plants: Passioura 1972), evolutionary cost–benefit considerations suggest a minimum biomass of water-absorbing and transpiring tissue which is appropriate for even a reductio ad (almost) adsurdum quantity of water-conducting tissue in the form of a single file of narrow elements. Similar considerations (apart from embolism, etc.) relate to the miniaturization of the phloem. A third factor which may dictate a lower limit on seedling size relates to the functioning of the stomata–intercellular gas space system. Functioning of the stomata–intercellular gas space system requires as an absolute minimum two cell layers, corresponding to the upper and lower epidermis of a planar system, with stomata in one or both of the layers and the two cell layers separated at least in part by a cell- and liquid-free void permitting gas diffusion. Such a minimal system demands photosynthetically active epidermal cells (non-stomatal as well as guard cells!) with free gas exchange between these cells and the intercellular gas spaces and restricted gas exchange between these cells and the atmosphere via the cuticle and epicuticular wax layer. Such structures are approached today in parts of the frond of some species of the fern Adiantum (Wylie 1948). In other extant vascular plant sporophytes the photosynthetic structures are more bulky, e.g. planar leaves with one or more layers of photosynthetic mesophyll between the epidermal layers, and cylindrical axes with one or more photosynthetic mesophyll layers under the epidermis. In these organisms photosynthetic activity in the epidermis is usually restricted to the stomatal guard cells. The possibilities of miniaturizing this supracellular homoiohydrically functional apparatus must relate to decreasing cell size rather than decreasing the number of cell layers. We deal in turn with each of these three possible constraints on the extent to which a photosynthetically self-sufficient, homoiohydrically functional seedling can be miniaturized. Dealing first with the height requirement to span the soil–atmosphere interface, we propose 5 mm as the minimum height (see Burslem, Turner & Grubb 1994; Metcalfe & Grubb 1997) which can at once gain access to water in the top few mm of soil and have a few mm of shoot in the air which can harvest photons (perhaps in rather short supply unless the seed is germinating in a relatively unvegetated soil, as is likely for very small-seeded ruderals) and CO2 (perhaps present at up to twice the bulk atmosphere concentration in the boundary layer where soil respiratory CO2 is incompletely dissipated). While transpiration may be restricted by higher relative humidity in the boundary layer near the soil, this effect would be offset if the seed is germinated in a high photon flux density environment in relatively unvegetated soil. This height of 5 mm we have assumed puts our hypothetical seedling in the size range of smaller specimens of sporophytes of such (probably) homoiohydric early vascular plants as Cooksonia (Kenrick & Crane 1997). The second point to be addressed is that of the minimum size of a functional endohydric water conductivity apparatus (in this case xylem). The quantitative key to this is the Hagen–Poiseuille equation ( eqn 1). where Jv is volume flux along a cylindrical conduit (m3 of water passing a certain point per second), r is radius of the conduit (m), l is length of the conduit (m), P is the pressure difference between the ends of the conduit (Pa) and η is the dynamic viscosity of the solution flowing along the conduit (Pa s–1). Our use of this equation assumes a cylindrical seedling 5 mm high (justified above) and 25 μm in radius (justified later as the minimum size of a homoiohydric photosynthetic axis). Jv is then estimated from the computed surface area of an assumed 2·5 mm long aerial portion (3·93 10–7 m2), with an internal photosynthetic area exposed to a gas phase which is twice the external area (i.e. 7·85 10–7 m2), a photosynthetic rate of 2 μmol CO2 per m2 internal area exposed to a gas phase per second and a transpiratory loss for C3 physiology of 666 mol per H2O per mol CO2 fixed (1000 g water lost per g C fixed). This gives value of 1·9 10–14 m3 axis–1 s–1 (see Raven 1984a, 1993, 1997a). We now assume that ΔP cannot exceed the Ψw at which the photosynthetic rate of vascular plants becomes restricted by too low a Ψw; assuming soil Ψw = 0 the shoot Ψw is some – 2 MPa (Raven 1993), so ΔP is taken as 2 106 Pa, l is taken as 2·5 10–3 m, i.e. assuming that the average water molecule has to move in the xylem for half of the length of the plant axis in the soil and half of the length of the plant axis in the air. With a η value of 10–3 Pa s–1 at 20 °C, the required value of r for a single conduit is 0·5 μm; this is likely to be an underestimate because Hagen–Poiseuille estimates of Jv values are generally two or three times higher than measured Jv values for xylem. Certainly the value for r is almost an order of magnitude less than the 4 μm of the smallest known xylem conduits (Raven 1977, 1984a, 1993), so that ‘real’ xylem, even a single file of elements 4 μm in radius, substantially overprovides longitudinal water transport capacity for our ‘model’ seedling. Indeed, based on a conductivity of parenchyma tissue (Raven 1977, 1984a, 1993) of 10–13 m2 s–1 Pa–1[i.e. a flux of 10–13 m3 (m2 cross-sectional area)–1 s–1 for a Ψw gradient of 1 Pa m–1], the 2·5 mm mean water transport distance of our axis could support a flux (Jv) through its 1·96 10–9 m2 of cross-sectional area of 1·57 10–13 m3 s–1, i.e. almost 10 times the actual flux. Accordingly, transport through parenchyma tissue could support the required Jv, and the xylem is at this stage of growth superfluous to requirements. However, xylem will be needed a little later in growth and it may be useful that xylem continuity along the axis is put in place ab initio so that it can be used, and built on, later in growth. However, we note that plant morphogenesis would permit the differentiation of xylem in a parenchymatous cylinder of seedling axis (Byrne, Pesacreta & Fox 1977; Byrne, Byrne & Emmitt 1982; Fukuda 1996; McCann 1997) when water conduction through parenchyma becomes inadequate. The seedling can thus cope with water fluxes without xylem early on but may need xylem differentiation at this early stage to provide conducting capacity for later requirements. The third and final point concerning the minimum size of a homoiohydric seedling relates to the minimum size of the stomata–intercellular gas space system. Here we have the minimum requirement of two photosynthetic cell layers separated, at least at some points, by a gas space system. Maximal effectiveness of the system in supplying CO2 to plastids requires that the two cell layers are not cuboid cells completely joined on four sides to adjacent cells where air spaces penetrate (Raven 1977). This is certainly the case for systems made up of cells of a ‘normal’ size, i.e. cells of maximum dimension 10–20 μm. It would be possible to circumvent the need for these clefts in the epidermal epithelium, as well as making the whole system smaller, by dealing with smaller cells. Raven (1984b, 1986, 1991a, 1996a) considers the costs and benefits of different sizes of photosynthetic cells with respect to constructional costs and their effectiveness in harvesting CO2 and photons, as well as other ecologically related attributes, for unicellular aquatic organisms. The conclusion here was that the most effective return on capital (energy, C, N, P, etc.) invested in photon-harvesting machinery and in CO2-assimilatory machinery was to distribute the machinery among a larger number of smaller cells than few, larger cells. The arguments related to self-shading of chromophores in the case of photon harvesting, and the thickness of external diffusion boundary layers and the internal diffusion pathlength in the case of CO2 (Raven 1984b, 1986, 1991b, 1996a). Parkhurst (1994) points out that minimal limitation of photosynthesis by restricted CO2 diffusion in the gas phase in photosynthetic tissues requires thin leaves (or small-diameter axes) and closely spaced stomata. These desiderata could most readily be achieved by small cells in relatively small numbers. Raven (1984b, 1987, 1997b) points out that these benefits could in large part be achieved by vacuolation of small cells; a possible exception is CO2 diffusion within gas spaces (Parkhurst 1994). This is at least as relevant to the homoiohydric seedlings we are dealing with here as it is to the unicellular organisms considered by Raven (1984b, 1986, 1987, 1991a, 1996a, 1997b). Thus, vacuolation permits the spreading of N and P more optimally in photon and CO2 harvesting, albeit with a cost in extra energy and C inputs related to the greater C and energy costs of wall synthesis per unit N and P in a vacuolate than in a non-vacuolate cell. In terms of photosynthetic organisms as whole, and our homoiohydric seedlings in particular, vacuolation increases the effectiveness of use of dry matter, C, N, P and energy resources in resource acquisition. Thus the option of small and of necessity non-vacuolate (Raven 1984b, 1987, 1997b) cells in miniaturizing homoiohydric seedlings is perhaps not evolutionarily viable because the small size is not necessarily economizing on seed resources because it involves a lower water (obtained from the soil) mass:dry mass ratio. Thus, the plant with small, non-vacuolate cells is smaller per unit dry mass invested and less competent at resource harvesting. We note that naturally or artificially bonsai plants have smaller organs but not smaller (or less vacuolate) cells (Korner, Menendez-Riedl & John 1989; cf. Parkhurst 1994). Accepting that the photosynthetic apparatus is made up of vacuolate cells c. 10 μm in diameter, the minimum size of a photosynthetic organ is a laminar structure 20 μm thick, noting that the occurrence of spherical cells allows for the presence of air spaces (Raven 1996b, 1998) although the specification of a spherical shape presents problems in terms of cuticular continuity. If an internal area exposed to gas spaces equal to more than the total external surface area is to occur, thus increasing the return on resources in terms of CO2 harvesting by decreasing the fraction of cell area abutting non-productively (for CO2 acquisition) onto the external cuticle, then an extra cell layer is needed regardless of whether the epidermal cells retain full photosynthetic capacity. Such an increase in cell number (of a given size) per unit external area also increases the potential for intercellular reflectance to maximize photon absorption per unit photosynthetic apparatus in a given light field (De Lucia et al. 1996; Raven 1996a,b). This would increase the thickness of a laminar photosynthetic apparatus to 30 μm. For a cylindrical photosynthetic apparatus (e.g. a hypocotyl) doubling of the external area in terms of the area exposed to the intercellular gas space system, together with the presence of at least one file of xylem elements with associated parenchyma (see Raven 1991b, 1994 for the uses of such parenchyma sheaths) requires a radius of the organ of 25 μm if the epidermal/mesophyll cells are to be essentially isodiametric with a diameter of 10 μm. These three considerations of height (length) needed to span the soil–atmosphere interface, minimum functional size of xylem and minimum functional size of the intercellular gas space–stomata–cuticle system, give similar results in defining the minimal size of a homoiohydrically functional, photosynthetically competent seedling. This minimal size of a homoiohydric organism can be compared with what can be produced from the quantity of reserves in the smallest known seeds yielding non-mycotrophic, non-parasitic (on other plants) seedlings. The volume of the 5 mm long, 25 μm radius cylinder is 9·82 10–12 m3; if 5% of the volume is occupied by gas spaces (Raven 1996b) and the density of the remaining tissue is 1·05 mg m–3 then the fresh mass of tissue is 9·8 μg. Granted a minimum seed dry mass of 5 μg (Metcalfe & Grubb 1995), and a content of useable food resources (endosperm plus cotyledons) plus embryonic axis of 4 μg (Fig. 4·28 and 4·29 of Moorby 1981; Fig. 1·8 and 1·9 of Freney 1985), then 2 μg dry mass of seedling could be produced granted a 2 μg of dry mass loss in maintenance and growth-related respiration in the seed plus seedling system (see heterotrophic embryo growth entries in Table 2 of Raven 1976) and in loss of soluble matter upon imbibition, and assuming no photosynthetic input of dry matter. This 2 μg of dry matter translates into 12 μg fresh mass of seedling, granted a mean fresh:dry mass ratio of 6 (with rather lower values likely in the shoot and higher values in the root, taking into account the necessity for a significant degree of vacuolation (see above and Raven 1987). This 12 μg fresh mass produced from 4 μg dry mass of useable seed reserves plus embryo is rather more than the 9·8 μg fresh (1·6 μg dry) mass computed for the smallest homoiohydric organism, so that a 5 μg seed could produce a homoiohydrically competent seedling. It is not easy to conceive of a homoiohydrically functional photosynthetically competent seedling which is significantly smaller than the one posited either for reasons of achieved homoiohydry (height; multicellularity of photosynthetic apparatus with gas spaces, stomata and cuticle), or of homoiohydry of large derivatives of the seedling (even if there is no necessity for functional xylem in 50 μm diameter, 5 mm high plants). Similar arguments to those developed for the multicellularity of the shoot could be developed for the radical. Indeed, the radius of roots of extant (and fossil) plants never seems to be less than 25 μm (the value assumed for our 5 mm long cylindrical axis), with any absorptive or anchoring organs of smaller radius than 25 μm taking the form of aseptate evaginations of epidermal cells (root hairs) or septate or aseptate fungal hyphae (mycorrhizas): see Fitter (1987). Accordingly, it seems possible that a constraint on the minimum size of seeds is the minimum quantity of food reserve plus embryonic axis which is necessary to produce a seedling which is both photosynthetically self-sufficient and homoiohydrically functional. This does not, of course, mean that there are no other physiological or ecological constraints on the minimum size of seeds producing such seedlings without any energy and C input other than from the seed. However, the occurrence of significantly smaller seeds for at least some plants whose seedlings (and often the older stages of the plant) obtain C and energy by parasitizing their mycorrhizal fungi, or another (photosynthetic) plant, suggests that the constraint on minimal seed size for plants not receiving such subsidies cannot be related to intrinsic features of embryogenesis or of the extent of development of embryos prior to entering dormancy. It is of interest that some ‘anomalous’ (for a plant) modes of nutrition do not lend themselves to enabling the plant to produce seeds less than 5 μg in dry mass. One such nutritional mode is that of insectivory; even the capture of a small insect by the small-leaved Western Australian sundew Drosera menziesii requires the production of a trap (leaf) weighing more than 10 μg dry mass (Pate & Dixon 1981; Morley & Toelken 1983)! However, this may not be true of the recently discovered protozoa-trapping specialism of the bladderwort Genlisea (Barthlott et al. 1998). The minimum seed size of 5 μg dry mass for non-mycotrophic, non-parasitic terrestrial plants involves slow-growing seedlings establishing in habitats which may have infertile soil or deep shade, but which have a very reliable water supply (Grubb 1996; Metcalfe, Grubb & Turner 1998). The earliest known seeds are substantially larger than this minimum 5 μg dry mass (Harper et al. 1970; Chaloner & Sheerin 1981; Taylor & Taylor 1993). Finally in this section we return to the question of the speed with which a seedling can ‘bridge the gap’ between a dormant and frequently desiccation-tolerant seed and a photosynthetically self-sufficient, homoiohydrically competent seedling. Although the early stages of seedling growth may not be much less demanding of soil water than seed imbibition and germination per se, in terms of the Ψw or soil water content required, the early protrusion of the radical enables a greater volume of soil to be explored. Furthermore, the epicotyl (with or without photosynthetic cotyledons) would, if optimization principles were followed, not begin to lose water via open stomata until photosynthetic competence had been established. However, a ‘chloroembryo’ which continued net photosynthesis after germination until it became photosynthetically independent might lose more water in transpiration than would a seedling which was developing photosynthesis but with limited gas exchange with the atmosphere. This, and other aspects of the suggestions made here, is subjectable to experimental test. The evolution of homoiohydry can be deduced from the presence in fossil plants of the attributes listed earlier, i.e. roots (or their functional equivalent), xylem (or its functional equivalent in permitting endohydry, i.e. the ability to transport all of the water lost in transpiration in specialized internal apoplasmic water mass flow pathways), cuticle, stomata and intercellular gas spaces. Clearly a good level of structural preservation is needed to make sure that the absence of one or more of the requirements for homoiohydry is unlikely to be a result of poor preservation. In the absence of the full suite of homoiohydric features in a fossil plant the organism must be assigned, at least provisionally, to the poikilohydric category, i.e. is a plant with little control over its degree of hydration with varying soil water supply and evaporative demand by the atmosphere. Organisms with some but not all homoiohydric features exist today among the gametophytes and sporophytes of various bryophytes (Raven 1993). An example is provided by the gametophytes of polytrichaceous mosses, with an endohydric water-conducting system (hydrome) connecting a ‘root’-like rhizome/rhizoid system below ground to a photosynthetic apparatus in ‘solid’ tissue covered by a cuticle but lacking stomata and developmentally defined intercellular gas spaces (they have longitudinal photosynthetic lamellae on the stratose leaves which lack full physiological capacities of homoiohydry). A further example are the gametophytes of marchantiaceous liverworts with intercellular gas spaces (perhaps with different developmental origins to those of embryophyte sporophytes) and cuticle but with epidermal pores which do not function as stomata and no endohydric-conducting system (Raven 1993). A final example is provided by the sporophytes of these mosses and those of hornworts, with a capsule possessing cuticle, stomata and intercellular gas spaces, but connected to a ‘placental’ (parasitic) foot embedded in poikilohydric gametophyte tissue by a seta lacking an endohydric conducting system. Among vascular plants the subterranean mycotrophic gametophytes of a tetraploid race of Psilotum nudum has vascular tissue and cuticle (Bierhorst 1971; Whittier & Petersen 1995). These present-day examples of the presence of only a limited number of the five essentials for complete homoiohydry alert us to the strong likelihood of finding such organisms in the fossil record without their necessarily being intermediates on the evolutionary pathway to complete homoiohydry. Thus, the occurrence of some of the homoiohydric attributes may confer evolutionary advantage to some land plants today and, presumably, in the past. Turning to the evidence for complete homoiohydry, or for the occurrence of certain homoiohydric features, in the fossil record, the first evidence for the occurrence of the full range of homoiohydric features comes from sporophytes of plants from Upper Silurian strata such as Baragwanathia (a lycophyte) and Cooksonia (a rhyniophyte). The Lower Devonian sporophytes Rhynia, Horneophyton and Asteroxylon are also homoiohydric on the basis of structural features, while the Lower Devonian sporophyte Aglaophyton major (formerly Rhynia major) does not have true xylem but combines an endohydric conducting system with an underground axial system functioning like roots and cuticle, stomata and intercellular gas spaces in the aerial axis (see Raven 1984a; Raven 1993). Evidence provided by Remy & Remy (1980a,b) and Remy & Hass (1991a,b) on fossils from (as are Asteroxylon, Aglaophyton, Horneophyton and Rhynia) the Lower Devonian Rhynie Chert deposit and from other Lower Devonian strata (Remy 1982) shows the occurrence of gametophytes (see Taylor & Taylor 1993; Kenrick & Crane 1997) with homoiohydric characteristics. That Kidstonophyton, Langiophyton and Lyonophyton are embryophytic gametophytes is suggested by the presence of antheridia and archegonia. These gametophytes have stomata and (apparently) endohydric-conducting tissue; the occurrence of intercellular gas spaces, cuticle and structures which could take up water from soil is less well established. It is believed on anatomical grounds that these gametophytes are phases in the life cycle of the sporophytes of vascular (Asteroxylon, Horneophyton, Rhynia) or vascular-like (Aglaophyton) plants found in the same strata (Taylor & Taylor 1993). However, these fossils could be the gametophytes of bryophytes (see Taylor & Taylor 1993) with similarities to the extant order Marchantiales, which have cuticle and intercellular gas spaces but pores rather than stomata sensu stricto, and no real endohydric conducting sy