Understanding the mechanisms of P availability to plants is critical for sound management of both croplands and forests. Phosphorus availability often limits forest growth, especially on highly leached soils of old landscapes, and where nitrogen fixers reduce N limitations. Limitation by P is also commonly observed in heathlands as well as in upland and montane grazing environments. The importance of P for crop production is obvious from the world-wide practice of P fertilisation. Maintaining adequate P levels is often difficult. Unlike N, P does not enter the ecosystem in any significant amount from the atmosphere, and both total and available stores decrease steadily with time (Walker and Syers, 1976).
Studies of P cycling and availability have posed a challenge to agronomists and ecologists for many years because P exists in soils and sediments in many different physico-chemical forms, and it is involved in a myriad of biological processes. The cycling of P can be controlled by inorganic chemical reactions, but in many systems, the turnover of organic P controls the availability of P to plants through organic matter decomposition, or the release of P from microbial stores (Chapin et al., 1978; Harrison, 1985; Walbridge, 1991). Phosphate sorption by various soil constituents is a dominant process maintaining soil solution P at very low levels, and it has proven difficult to measure the amount of P in a soil that is available for plant uptake. Whereas important pools and fluxes of nitrogen are readily measured in soils (e.g., "soluble" inorganic N, rate of N mineralisation), the different physico-chemical forms of P are difficult to categorise or to define operationally, thus limiting our ability to develop a simple definition of bioavailable P.
Perhaps a more serious problem for the study of plant-available P is the observation that plants differ in their abilities to grow in highly infertile soils or in soils with a high phosphate sorption capacity. Clearly, plants have differing physiological mechanisms and strategies for acquiring and/or retaining adequate P from the soil solution. There is also increasing evidence that certain plants may be able to obtain soil P fractions held in specific physico-chemical forms that are not available to all plants (Ae et al., 1990; Lajtha and Schlesinger, 1988b; Bolan et al., 1984). Although there is still debate over whether a mycorrhizal association allows a plant to use forms of soil P that are unavailable to nonmycorrhizal plants, it appears that certain species native to soils high in metal oxides or carbonates, or simply very low in P fertility have specific adaptations that might allow them either to solubilise "unavailable" soil P fractions or to use P more efficiently. In addition, plants may employ different nutrient conservation strategies that might allow them to grow in low-P soils. Without a mechanistic understanding of the physiological processes that underlie these differences, predictive models of plant growth in P-limited environments will not be possible. However, such an understanding is necessary if we are to continue sustainable agriculture in soils with a high P sorption capacity or with reduced P fertilisation (Hands et al., Ch. 10), or to select crops that will survive under such conditions.
In this chapter we will discuss strategies and physiological adaptations of plants to low P availability. This includes mechanisms to exploit soil P pools in various physico-chemical forms, as well as strategies for P conservation within whole plants, tissues or genets.
Controls on Phosphorus in the Soil Solution
Soil solution P is usually quite low due to complex interactions of phosphate with various soil components. Thus plants must either employ mechanisms to increase the solubility or availability of these components or else rely on diffusion. Sorption is considered to be the most important process controlling P availability in soils. The dominant reactions are those with Fe-Al oxides and Ca-carbonates. Since sorption is to some degree reversible, sorbed P is a source of plant-available P either immediately or over a longer term.
The reaction between phosphate and reactive surface Fe and Al oxides is typical of acid soils. It is highly dependent on the crystallinity and surface area of the oxides present in soils, and thus on the mineralogy of the clays in soil. Experiments on the reversibility of P sorption have shown that the plot of sorbed P against concentration is not retraced during desorption (Barrow, 1983b), suggesting that sorbed P undergoes further transformations with time (Parfitt et al., 1989), perhaps also changing its availability for exchange into the soil solution and thus for plant uptake. Such processes may involve recrystallisation (Barrow, 1983b), solid or liquid state diffusion towards the centers of sorbing components, solid phase migration of phosphate into the interior of clay crystals, particularly along lines of crystal defects (Norrish and Rosser, 1983), or multiple P pools which are not in immediate exchange with the solution (Fardeau and Jappe, 1980) or which have differing affinities to P (Kanabo et al., 1978). The rate and extent of reversibility of these reactions is not well known, although these pools play an important role in P availability, yet perhaps at rates and at levels that are not sufficiently great to make biologically significant contributions to the soil available pool. Organically bound P may also undergo precipitation and fixation reactions with Fe and Al. Much of the secondary Al- and Fe- bound P remains unavailable to plants, as does the P associated with organic matter of a molecular weight >50,000 (Goh and Williams 1982).
Mechanisms of P sorption and the strength and reversibility of sorption in medium to high pH systems can differ substantially. Adsorption of P onto CaCO3 and the co-precipitation of Ca-P minerals is considered to dominate in both alkaline aquatic and terrestrial systems where carbonates are present (Cole and Olsen, 1959; Lajtha and Bloomer, 1988). However, calcareous soils may still have significant levels of Fe and Al oxides, either as discrete components or as coatings on other soil particles, and thus phosphate sorption may be controlled by the presence of metal oxides as in more acidic soils. Several P adsorption studies in calcareous soils derived from limestone have found stronger relationships between P adsorption capacity and hydrous oxides of Fe and Al than with soil CaCO3 content (Holford and Mattingly, 1975a; Solis and Torrent 1989; Ryan et al., 1985). Using optical microscopy and electron microprobe analysis, (Hamad et al., 1992) also found evidence that separate phase Fe oxide was the major contributor to P sorption in calcareous Sudanese soils, followed by Fe coatings, with little contribution from uncoated CaCO3. This might suggest that the controlling sorption processes across all soil types are functionally quite similar, and thus specific adaptations to specific soil types for P acquisition are unlikely. In arid soils of New Mexico, however, Lajtha and Bloomer (1988) found a stronger relationship between P sorption and soil CaCO3 content than with Fe and Al oxides and hydroxides. Lajtha and Schlesinger (1988a) also found that total P profiles with soil depth more closely matched the depth-distribution of total CaCO3, rather than total or dithionite-extractable Fe or Al profiles. Since the CaCO3 distribution was closely related to soil-forming processes, the authors hypothesised that these results pointed out the importance of reactive carbonate surfaces in stabilising soil P.
In soils derived from limestone parent material, weathering decreases total CaCO3 content but increases specific surface area of the now smaller CaCO3 particles (Holford and Mattingly, 1975b). Thus any relationship between CaCO3 and sorption is obscured by these opposing trends, since total carbonate decreases but carbonate surface area increases with increased soil weathering. In semiarid soils where CaCO3 is continuously input from aeolian dust, an inverse relationship between CaCO3 surface area and content might not be expected. It is quite likely for soils in semiarid environments to undergo periodic cyclic dissolution and reprecipitation to form small, high-surface-area CaCO3 grains that would be highly reactive.
In general, neutral to slightly alkaline soils have shown higher soil solution P levels than acidic soils dominated by oxide-rich clays, suggesting that sorption processes may not have such a marked effect on P bioavailability in these systems. However, specific adaptations to low P availability are observed in calcareous systems, and studies have shown that plants may vary in their ability to extract P from low-P, calcareous soils. For example, Lajtha and Schlesinger (1988b) noted that effects of experimental additions of CaCO3 to desert shrub species were species specific; Larrea tridentata responded with decreased tissue P and increased root:shoot ratios, while Parthenium incanum, a shrub restricted to eroded, high carbonate soils did not show decreased tissue P in response to increased carbonate content. Yet while many of the physiological mechanisms observed for efficient P extraction are seen across all soil types, others appear to be specific to specific physico-chemical soil characteristics.
Plant adaptations
Plant adaptations to P stress can be loosely classified into strategies to maximise P uptake and strategies for the efficient use and conservation of P once it is obtained (Table 1). Many mechanisms to maximise soil P uptake are general responses to any nutrient stress and may also be found when other nutrients, such as nitrogen, are limiting. Increases in root:shoot ratios, for example, are observed either when one element is strongly growth-limiting or when ratios among specific soil resources are unusually skewed (Gulmon and Chu, 1981; Lajtha and Klein, 1988).
Table 1. Strategies for the uptake and acquisition of P, and for conservation of P within plants. Not all species will employ all of the possible strategies.
|
Conservation / use |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Root uptake capacity
There is evidence that root architecture and rooting depth may influence P extraction from soils. Studies using 32P in soils have suggested that most temperate and tropical tree species take up most of their P from very near the surface (Harrison et al. 1988). However, plants may adjust rooting depth and architecture in the face of nutrient competition. Goodman and Collison (1982) demonstrated, using 32P, that a grass and a clover took up P from near the surface when grown in monoculture conditions. However, when grown in a mixture, clover rooted deeper in the soil, due to either water or nutrient stress, and took up P from lower soil horizons. This clover/grass mixture was more productive than single monocultures, presumably due to exploitation of a greater soil volume. This study has clear implications for mixed cropping systems, and would suggest that in mixture, species or varieties with possibilities for differences in rooting depths could exploit soil volume better than could single monocultures.
Species with low relative growth rates appear to be more tolerant of low nutrient supply than fast-growing species (White, 1972; Clarkson, 1967; Rorison, 1968) but are less responsive to increases in nutrient availability (Nassery, 1970; Chapin et al., 1982; Veerkamp and Kuiper, 1982). Thus low growth rate appears to be an effective mechanism to cope with low fertility, or low P soils. Less is known about the ability of roots to adapt nutrient uptake capacities to low P availability. Results of root nutrient uptake studies are few, and have produced inconsistent results. Plants from nutrient-poor habitats have shown either lower capacities for phosphate absorption than plants from fertile habitats (Rorison, 1968; Chapin, 1980; Chapin et al., 1982; Chapin, 1983; Chapin et al., 1986a) or similar capacities (Veerkamp and Kuiper, 1982), or else uptake capacities were greatest in populations with the highest genetic potentials for growth, and were not necessarily related to habitat fertility (Chapin and Oechel, 1983). Many of these experiments on nutrient uptake kinetics are confounded by the ecologically plastic character of P uptake rates, which can depend as much on plant nutrient status as on nutrient availability in the immediate growth environment (Harrison and Helliwell, 1979; Atwell et al., 1980). Roots increase their capacity to absorb limiting nutrients as reserves of that nutrient decline in the plant (Chapin, 1987). Chapin et al. (1986a) hypothesised that a high ion uptake capacity would be an asset for root competition in infertile soils only in the case of mobile ions such as potassium and nitrate for which diffusion in soil is relatively rapid. For immobile ions, such as phosphate and ammonium, diffusion would be the rate-limiting step in the uptake process. They suggested that there should be a selection for high phosphate uptake rate in plants adapted to fertile soils, but not in plants adapted to infertile soils because diffusion limitation would override variation in uptake capacity. In a similar study, Lajtha (1994) saw no relationship between growth rate or habitat and root P uptake rate in a suite of eastern deciduous tree seedlings. Thus, plants adapted to low P soils are not likely to have evolved mechanisms for efficient uptake, whereas changes in root architecture or more efficient soil exploration would be expected, and have indeed been observed.
It should be noted, however, that a relationship between root uptake rate and any other plant characteristic can only develop if P is the primary factor limiting plant growth in the ecosystem or experimental system. Harrison and Helliwell (1979) argued that P uptake rate is determined primarily by the demand for P within the plant relative to the availability in the soil. A measure of P uptake rate might therefore offer an index of plant P availability. These authors described a plant bioassay technique for directly comparing P availability in soils. Although this technique produces an index of P bioavailability rather than a quantitative measure of an "available P" soil pool, it has the advantage of directly using plants to provide a relative measure of P limitation. If root uptake of a nutrient increases when plant demand for the nutrient increases, it may be possible to introduce labelled P into the (soil) solution surrounding the roots in vitro, and then the amount of 32P taken up by a plant's roots should be dependent on the bioavailability of P in the plant's environment. Thus this bioassay defines "available P" in relation to plant demand for P. This demand can be influenced by many factors, such as the relative availabilities of other nutrients such as N or K (Jones et al., 1991), or even climatic conditions that affect plant production potential. A direct bioassay technique has important advantages over soil measurements of P bioavailability, particularly since plants may have differing abilities to take up P fractions of different chemical nature and availability. Only a bioassay technique will detect this difference in P acquisition, whereas soil extracts or sorption/desorption studies will not.
Mycorrhizae and root exudates
Many authors have used 32P isotope exchange experiments to test the hypothesis that mycorrhizal plants can use chemical forms of P that are unavailable to nonmycorrhizal plants (Sanders and Tinker, 1971; Hayman and Mosse, 1972; Powell 1975). Since the resulting specific activities of P in mycorrhizal and nonmycorrhizal plants in these experiments have been shown to be similar, authors have generally concluded that mycorrhizal associations do not afford greater access to sorbed or "fixed" P. It has therefore been argued that the advantage of mycorrhizae is one of increased soil exploration, similar to increased root:shoot ratios. However, several 32P exchange techniques have been called into question, and 32P exchangeable P may not correspond entirely to plant available P. In greenhouse trials, Bolan et al. (1984) added FeOH to pots containing either nonmycorrhizal subterranean clover or clover inoculated with VA mycorrhizae, and found decreased P uptake by the nonmycorrhizal plants but no effect on growth of the mycorrhizal plants. The FeOH additions had no effect on the specific activity of P in any of the plant groups. From these data the authors concluded that mycorrhizal associations allowed the plant to use P sorbed by iron oxides that was unavailable to nonmycorrhizal plants. There is also limited evidence that mycorrhizal fungi may be able to utilise organic P sources more effectively (Jayachandran et al., 1992), although this point is debated in the literature. Ectotrophic mycorrhizal fungi have varying abilities to produce phosphatases and hence to utilise organic P sources (Dighton, 1983).
There is evidence that different mycorrhizal types or species may have differing abilities to scavenge P from soils. Dighton et al. (1990) demonstrated that different species of ectomycorrhizal fungi had different abilities to effect phosphate uptake into host birch trees. Similarly, Dighton and Coleman (1992) demonstrated that different mycorrhizal associates of Rhododendron maximum L. can significantly affect the host plant's P status. This, in turn, would suggest that it might be possible to select for efficient ectomycorrhizal species and isolates for inoculation programmes (Dighton et al., 1993).
Mycorrhizal associations appear in all soil types, from highly acidic, low-P soils to desert soils high in free available carbonates. However, since there is strong evidence that mycorrhizal associations aid plant water balance in calcareous desert or grassland soils (Allen and Allen, 1986), it is difficult to conclusively assign a cause-and-effect relationship between mycorrhizae and improved P status in these systems.
Even without mycorrhizal associations, pigeon pea (Cajanus cajan), which is commonly intercropped with cereals under low fertilizer input conditions, appears to have root exudates that chelate Fe3+ and thus can utilise P bound by Fe oxides (Ae et al., 1990). This P can then be available to crop plants nearby; thus differential use of soil P by different species has strong implications for soil fertility and agriculture where P is a strongly limiting factor. This result leads to the possibility that plants with the ability to solubilise "unavailable" P can be sought and used in agricultural systems that show a P limitation to production. Even if such plants are themselves not crop plants, they could be used in alley cropping or physically grown between crop species, or else could be chronologically intercropped, as is commonly done for such species as lucerne. Species such as Inga edulis are potentially useful intercropping species, for example; this species appears to be able to acquire and cycle P in very high amounts on poor ultisolic soils of the humid tropics (Hands et al., Ch. 10). Unfortunately, such species may also become noxious weeds, possibly because of their competitive advantage. There is evidence that bracken fern (Pteridium aquilinum) has a high capacity to mobilise P through the production of root exudates (Mitchell, 1973); this species is thus able to spread rapidly over agricultural land, reducing its use for grazing.
The occurrence of oxalic acid in natural systems, synthesised either by plants or by microorganisms such as fungi, and its role in mineral weathering has been well documented (Jurinak et al., 1986; Knight et al., 1992). Ca-oxalate crystals have been observed in both the ectomycorrhizal mantle of pine roots and in external hyphal mats in acid soils of the Pacific Northwest (Cromack et al., 1979). Graustein et al., (1977) suggested that oxalate associated with pine ectomycorrhizae could cause the release of P from relatively insoluble Fe- and Al-phosphate minerals. Oxalate may form stable complexes with Al in soils or undergo ligand-exchange reactions with oxide-sorbed P, thus increasing P solubility in the soil solution (Fox and Comerford, 1992a). Similarly, Jurinak et al. (1986) found Ca-oxalate crystals associated with vesicular-arbuscular mycorrhizae (VAM) in a calcareous grassland, and suggested a similar role for oxalate in the release of P from Ca-P minerals such as hydroxyapatite. Indeed, Sollins et al. (1981) found that oxalate-producing mycorrhizae could solubilise Ca-bound P, whereas VA mycorrhizae that did not produce oxalates (such as Endogone) could not. If solution P is indeed controlled by Ca-P interactions and precipitation in calcareous soils, then the presence of Ca-oxalate should increase soil solution P by depressing Ca2+ ion activity and enhancing the dissolution of the P-controlling mineral (Knight et al., 1992). Lapeyrie et al., (1987) further suggested that oxalate synthesis by mycorrhizae was influenced by exogenous soil solution Ca2+ and HCO3-, and thus production of oxalate would be related to demand for solution P. The implications for soil P fertility are clear. For example, Allen and Allen (1988) demonstrated that Agropyron smithii had higher P uptake when inoculated with oxalate-forming mycorrhizae, or when grown in association with a nonmycotrophic weed, Salsola kali, that produces high levels of oxalate.
Conservation strategies
As with the various uptake mechanisms discussed, many of the P conservation strategies employed by different species of plants are general to any type of nutrient stress, rather than being specific to P stress. For example, prolonged tissue life is generally observed in nutrient poor ecosystems (Bloom et al., 1985), and will significantly increase the retention time of nutrients in the leaf biomass pool (Escudero et al., 1992). There are well-documented patterns of an increase in the proportion of evergreen species relative to deciduous species as habitat fertility decreases, both over local environmental gradients as well as globally (Beadle, 1966; Goldberg, 1982; Loveless, 1961; Monk, 1966). When relative growth rates are decreased by nutrient limitations, leaf longevity generally increases within a species (Bazzaz and Harper, 1977). However, several studies have found that leaf longevity decreases under strong nutrient stress within a species as nutrients from old leaves are preferentially translocated to support new plant growth (Osman and Milthorpe, 1971; Turner and Olsen, 1976; Reader, 1978). This translocation of nutrients from older to younger tissues appears to be a general response to nutrient limitation, seen in many tree species. For example, Fife and Nambiar (1982) showed that 83% of the P in the foliage of Pinus radiata had originated from internal translocation from older tissues. This translocation enables trees and other perennial plants to continue vertical growth and to compete with adjacent trees even in the face of nutrient limitation. Similarly, rhizomatous plants, and presumably those plants which propagate by developing a network of stolons, can transfer P from older regions to younger ones, effectively decoupling production from soil nutrient availability (Chapin et al., 1975; Greenway and Gunn, 1966; Hedley et al., 1985).
The amount of dry matter produced per unit tissue nutrient, which is one definition of nutrient use efficiency (NUE), generally increases under nutrient stress of any major nutrient (Chapin, 1987). Within a species, there is both field fertilisation and greenhouse experimental evidence that dry matter production per unit nutrient increases as nutrient stress increases (Birk and Vitousek, 1986; Ingestad, 1979). However, plants specifically adapted to low nutrient ecosystems do not always have greater NUE than plants adapted to high fertility sites, in part due to the low growth rates associated with plants from low fertility sites (Chapin, 1980). In a global compilation, Vitousek (1982) found that the efficiency of litter production per unit N was greater in coniferous forests than in temperate deciduous forests, which in turn was greater than in tropical forests. Coupled with the increase in leaf longevity of conifers, clearly N use efficiency is high in coniferous forests. However, patterns for P were quite different; the amount of dry matter produced per unit P was systematically higher in tropical forests, where P limitations to production are common.
The concentration of N and P in abscised leaves and litterfall is generally lower than that of live plant tissues. This nutrient resorption confers a second type of nutrient use efficiency on vegetation, and the efficiency of nutrient resorption varies significantly among species and among ecosystems. On average, about half of a leaf's P pool is resorbed prior to abscission, yet this varies from 0-90% (Chapin, 1987; Chapin and Kedrowski, 1983). Many authors have attempted to correlate nutrient resorption efficiency with plant nutrient status or soil fertility, with mixed results. Some studies have shown that nutrient resorption efficiency increased as nutrient stress increased, both in greenhouse and field studies (Miller et al., 1979; Shaver and Melillo, 1984; Small, 1972; Stachurski and Zimka, 1975; Turner, 1977; Vitousek, 1984), while others have found no such relationship or else have found reverse trends (Birk and Vitousek, 1986; Boerner, 1985; Chapin and Kedrowski, 1983; Lajtha, 1987; Lajtha and Klein, 1988; Lennon et al., 1985; Ostman and Weaver, 1982; Staaf, 1982; Walbridge, 1991). Perhaps part of this inconsistency can be attributed to the varying degrees of nutrient limitation found or imposed in these studies. Some authors have argued that total quantities of nutrients resorbed generally increase with increasing nutrient status in the plant, because more of the plant nutrient is then in a labile form, and thus percent resorption may be a misleading statistic.
Nutrients may also be closely recycled from fallen litter (Medina and Cuevas, 1989), and studies using 32P in soils have suggested that most temperate and tropical tree species take up most of their P from very near the surface (Harrison et al., 1988).
Along natural fertility gradients, several studies have suggested that nutrient resorption indeed increased as fertility decreased, in part due to changes in species composition (Zimka and Stachurski, 1976; Muller and Martin, 1984), suggesting large-scale ecosystem patterns of species' distributions. Other authors also have noted a tendency for species with higher inherent capabilities for nutrient resorption to dominate nutrient poor sites (Birk and Vitousek, 1986; Chapin et al., 1986b; McGraw and Chapin, 1989; Schlesinger et al., 1989; Schlesinger, 1991). In his compilation of data from forest ecosystems of the world, Vitousek (1982) found that litter N content generally declined in parallel with soil N availability, suggesting that N resorption efficiency increased. He found a similar pattern for P in litterfall in a compilation of data from tropical forests, where P limitations to production are common (Vitousek, 1984).
Resorption patterns of N have been studied more thoroughly than those for P, as N more commonly is the limiting nutrient in temperate terrestrial ecosystems. As a corollary, studies of P resorption in N-limited sites may not show predictable patterns, as was demonstrated in the studies of Vitousek (1982, 1984). However, some of the highest reported levels of P resorption are found in ecosystems with known P limitations or that show a response to P fertilisation, such as heathlands or Eucalyptus forests in Australia (Specht and Groves, 1966; Attiwill et al., 1978), Arctic tundra (Jonasson and Chapin, 1985), or acid, organic coastal wetland soils (Walbridge, 1991). Whether these global patterns will hold out upon further study is unclear. Just as we suggested that selection for plants with unique P acquisition strategies from soil pools might be possible in agriculture and agroecosystems, selection of plants with high P conservation capacities, that effectively decouple plant P status and soil P availability, might also be possible. Certainly this should be a priority for further research.
|
Phosphorus in the Global Environment. Edited by H. Tiessen © 1995 SCOPE. Published in 1995 by John Wiley & Sons Ltd. |
|
Last updated: 30.06.2001