Interpretation of bioassays in the study of interactions between soil organisms and plants: involvement of nutrient factors

Authors

  • S. R. Troelstra,

    Corresponding author
    1. Netherlands Institute of Ecology, Centre for Terrestrial Ecology, Department of Plant–Microorganism Interactions, PO Box 40, 6666 ZG Heteren, The Netherlands
    Search for more papers by this author
  • R. Wagenaar,

    1. Netherlands Institute of Ecology, Centre for Terrestrial Ecology, Department of Plant–Microorganism Interactions, PO Box 40, 6666 ZG Heteren, The Netherlands
    Search for more papers by this author
  • W. Smant,

    1. Netherlands Institute of Ecology, Centre for Terrestrial Ecology, Department of Plant–Microorganism Interactions, PO Box 40, 6666 ZG Heteren, The Netherlands
    Search for more papers by this author
  • B. A. M. Peters

    1. Netherlands Institute of Ecology, Centre for Terrestrial Ecology, Department of Plant–Microorganism Interactions, PO Box 40, 6666 ZG Heteren, The Netherlands
    Search for more papers by this author

Author for correspondence: S. R. Troelstra Tel: +31 26 479 1316 Fax: +31 26 472 3227 Email:troelstra@cto.nioo.knaw.nl

Summary

  •  Increased plant growth in sterilized soil is usually ascribed to the elimination of (often unidentified) soil-borne pathogens. Plant–soil bioassays are reported here for three dune soils and two plant species (Ammophila arenaria and Carex arenaria).
  •  Dynamics of plant growth, availability and uptake of nutrients were compared in sterilized (25 kGy gamma-irradiation) vs control soils.
  •  Plant growth, availability and acquisition of nutrients, for example P, even when provided in apparent excess, were significantly enhanced in gamma-irradiated calcareous dune sands. With A. arenaria, the positive sterilization effect occurred independently of initial plant dry mass. The addition of extracts of planted soils to A. arenaria growing in unsterilized sand caused an increase in root growth that could not be related to either nutrients or pathogens.
  •  Increased availability and acquisition of nutrients in sterilized soil may contribute to nonsterile : sterile ratios of plant growth that are < 1. Any ecological speculation involving the role of soil-borne biological factors should be based on fully validated plant–soil bioassays, which account for nutritional or other nonpathogen-related side-effects induced by soil sterilization.

Introduction

Over the last few decades, interest has increased in the role of soil pathogens in plant life histories and their effect on successional dynamics and diversity of plant species in (semi)natural vegetation (Jarosz & Davelos, 1995; Bever et al., 1997; Clay & Van der Putten, 1999; Olff et al., 2000; Packer & Clay, 2000; Van der Putten, 2000). Soil sterilization, the addition of selected soil biota to sterilized soil and reciprocal transplantation (growth of species A in rhizosphere soil of species B, and vice versa) are common techniques used in soil-pathogen studies. Dune ecosystems have served as a model system in the study of soil-borne plant pathogens using gamma-irradiation or biocide amendment of soil to demonstrate a soil’s pathogenic potential (Oremus & Otten, 1981; Maas et al., 1983; Van der Putten et al., 1988; Van der Putten & Troelstra, 1990; De Rooij-Van der Goes, 1995; Zoon, 1995). This and other research has suggested a role of harmful rhizosphere organisms in the natural decline of plant species and vegetation succession in coastal foredunes (Van der Putten et al., 1993; Van der Putten & Van der Stoel, 1998; Van der Putten, 2000). There are indications that each plant species in the vegetation zonation has its own specific pathosystem (Van der Putten et al., 1993) and that different plant species have adopted different escape strategies (De Rooij-Van der Goes et al., 1995; D’Hertefeldt & Van der Putten, 1998).

Unfortunately, experiments involving sterilization or related manipulations of soil usually correspond poorly to the dynamic field environment. Notwithstanding the apparent agreement between bioassays and many field observations in natural ecosystems (spatio-temporal patterns in vegetation; failure of replants to successfully establish), it often proves difficult to link directly the results for, for example marram grass to the existence and activity of specific pathogens or pathogen complexes (Van der Putten et al., 1990; De Rooij-Van der Goes, 1995; Kowalchuk et al., 1997; De Boer et al., 1998; Van der Putten & Van der Stoel, 1998; Van der Putten, 2000). Most bioassays using plant-soil systems start with small seedlings without precultivation and refer to a single harvest, usually after 6–8 wk. Obviously, such procedures do not provide insight into the dynamics of plant growth in sterilized vs nonsterilized soil over the preceding growth period. However, such experiments are often exclusively used to demonstrate a relatively low plant yield in nonsterilized as compared with γ-irradiated soil, resulting in a nonsterile : sterile (NS : S) ratio < 1. Although such observations may serve as a first hint to the presence of plant pathogens in the nonsterilized soil, the comparison of sterilized and nonsterile dune soils is still a ‘black box’ approach and much of the evidence for pathogens remains correlative or indirect.

An important implicit assumption in the comparison between control and treated sand is the absence of any (relative) nutrient limitation. In other words, since the nutrient supply is supposed to be equally adequate, differences in plant growth cannot be attributed to nutritional differences among treatments. However, optimal nutrient regimes for dune plants are often not known. Similarly, apart from its pattern of supply, other nutritional information (e.g. plant chemical composition) is often lacking. An exception is the study by Van der Putten et al. (1988), in which it was shown that growth inhibition of marram grass in control sand was most severe in the early stages of plant growth, coinciding with low apparent specific uptake rates of N, P and K. They suggested that the low nutrient uptake rates caused the pathogen-induced growth reductions, but this issue received little attention in subsequent studies.

The primary objective of this study was to test the hypothesis that, under the same nutrient regime, plants grow better in sterilized (gamma-irradiated) vs nonsterile soil as a result of increased nutrient availability. A second objective was to examine the effect of plant size on the outcome of the bioassay. Thirdly, to test the assumption that improved plant performance in sterilized soil might also be related to some nonnutritional stimulus, the effect of various bioassay soil extracts on plant growth was measured.

Materials and Methods

Soils

The experiments were conducted using root-zone topsoil (0–30 cm) collected from three coastal dune sites in the Netherlands. From north to south: (T) a site on the inland (53°25′ N, 5°24′ E) of the Wadden island Terschelling with a mixed marram grass-sand sedge vegetation; (V-1) a site (51°55′ N, 4°04′ E) on the island of Voorne with a cover of largely sand sedge; and (V-2) a dune ridge (51°51′ N, 4°04′ E) on Voorne, situated close to the Haringvlietdam, covered with mainly marram grass. Both locations containing marram grass showed vigorous growth of this species, although it was less abundant and more variable with respect to percentage cover at site T. Relevant soil characteristics are given in Table 1. Soil samples were always collected shortly before use. Soils were sieved (< 4 mm) and part of the retained roots and rhizomes were finely cut (< 1–2 cm) and reintroduced. Two sand portions were prepared, each provided with a basal nutrient supply, after which one was sterilized (25 kGy γ-irradiation; Gammaster, Ede, The Netherlands). Macronutrients added via the basal supply were (mmol pot−1): 1.25 Ca(NO3)2, 0.25 KH2PO4, 0.375 K2SO4, and 0.7 MgSO4; and micronutrients (µmol pot−1): 45 FeIIINa-EDTA, 45 MnCl2, 230 H3BO3, 4 ZnSO4, 1.6 CuSO4, and 2.6 Na2MoO4. In expt 1 (see below) the entire P supply had been mixed with the sand in solid form (0.9 mmol CaHPO4 pot−1) and the initial K2SO4 was increased to 0.5 mmol pot−1.

Table 1.  Selected soil characteristics (0–30 cm) of three coastal dune sites in the Netherlands
LocationpHCaCO3 (%)Organic C (µg g−1)Total N (µg g−1)NH4-Na (µg g−1)NO3-Na (µg g−1)Olsen-Pa (µg g−1)Particle size distribution (%)
H2OCaCl2< 150 µm150–300 µm> 300 µm
  • a

    , extractions of fresh soil; n.d., not determined.

Terschelling, T4.44.3032001701.40.10.61089 1
Voorne, V-18.27.54.938002401.42.51.23762 1
Voorne, V-28.67.73.81500< 100n.d.n.d.n.d. 86725

Plant cultivation and growing conditions

Seeds of marram grass (Ammophila arenaria (L.) Link) and sand sedge (Carex arenaria L.) were collected from the sites of soil sampling. Unless noted otherwise, germinated seedlings were pregrown for 3 wk in propagation trays with cone-shaped 30-ml cells filled with sterilized dune sand. The sand was saturated once with a complete nutrient solution. Plants were precultivated in the same soil as used for the main experiment. Following precultivation, plants were selected for the experiments in such a way that all treatments included similar classes of total leaf length, and usually referring to the same number of leaves per plant, at the narrowest possible range of classes. Regressions between total leaf length and shoot and root dry mass (70°C, 48 h) for the remaining plants were used to estimate dry mass of the 3-wk-old seedlings (n = 50–150; shoots: r2 > 0.60, P < 0.001; roots: r2 > 0.30, P < 0.001). Plants were bulked per plant part, ground (< 0.2 mm) and stored until further analysis (data for t = 0).

Unless noted otherwise, 1.5-l pots were filled with either sterilized or nonsterilized sand (equivalent to 1365 g d. wt) containing the basal nutrients. Pots were covered with perforated Al foil to minimize evaporation losses. One seedling (+ sand cone) was transplanted into each pot using a PVC cone-shaped planting gauge. Deionized water was added to raise the water content to 10% (w/w). All experiments were carried out in a greenhouse at 21/16°C (day/night) with additional lighting and a 16-h photoperiod. The lamps alone provided a minimum photon irradiance of 200 (expts 1 and 2) or 300–350 µmol m−2 s−1 (other experiments) at plant level. Pots were randomly arranged with regular rotation and rewatering to the initial moisture content with deionized water. Macronutrients were added at an increasing rate in wk 1, 3, 5 and 7: 0.25 × , 0.5 × , 0.75 ×  and 1 × the basal supply, respectively. Unless indicated otherwise, the following experiments always refer to sets of both sterilized and nonsterilized sand.

Experiment 1 (A. arenaria)

Seedlings (population Voorne) were transplanted into V-2 or T sand. Voorne plants in V-2 sand were harvested after 14, 27, 42 and 55 d, while those in T sand were harvested after 57 d (one final harvest). Treatments were replicated 15 times per harvest.

Experiment 2 (A. arenaria)

This experiment consisted of two time series for seedlings of population Terschelling in two sand types (V-2 and T). Ten replicates of each combination were harvested 14, 28, 42 and 56 d after transplanting.

Experiment 3 (C. arenaria)

Seedlings were grown in V-1 sand and harvested 14, 28, 42 and 56 d after transplanting. Treatments were replicated 15 times per harvest.

Experiment 4: effect of plant size

Four bioassays (a–d) were performed with A. arenaria (population Voorne) in V-2 sand with plants of increasing weight transplanted at the start of the experiments. Germinated seedlings (a), 3-wk-old seedlings (+ sand cone) (b), 7-wk-old plants (4 weeks growth in 1.5-l pots with sterilized sand) (c), and 11-wk-old plants (8 wk growth in 1.5-l pots with sterilized sand) (d) were used, respectively. For series a, seedlings were directly transferred, that is without precultivation, and initial dry mass was estimated from bulked samples of 150 seedlings. During the preparation of series c, 25 additional plants had been cultivated for 4 wk in sterile sand and harvested at the start of the experiment. Relative growth rate calculated for these plants was used to estimate initial dry mass of the experimental plants. For series d, a similar procedure was followed, but now the growth data of the b series provided the necessary parameters for estimating initial data. The roots of the plants for series c and d were washed free of soil and rinsed in deionized water. Plants were transferred to 3-l pots (c) or 9-l containers (d), filled with 3 and 9 kg sand, respectively, at 10% moisture (w/w). Basal nutrients and fertilization during the experiment were increased accordingly: × 2 and × 6, respectively. For transplanting, a hole (3 cm diam. × entire soil depth) was made in the centre of the pots and the roots, together with the removed sand, carefully washed into the hole with some deionized water. All series were harvested after 8 wk with 25 (a,b) or 15 (c,d) replicates per treatment.

Experiment 5: availability of P in Voorne (V-1) sand

Sterilized and nonsterilized V-1 sand was amended with KH2PO4 at rates corresponding with 0 × , 1 × and 2 × the basal supply, either with or without added carbon substrate (1 mg C g−1 as glucose + cellulose) to intensify immobilization. The sand (100-g portions) was 1% inoculated with sterilized sand from the trays with 3-wk-old seedlings (to simulate the effect of transplanting) and incubated at 20°C and 8% (w/w) moisture. NaHCO3 extractions (three replicate sand portions per treatment) were made after 1 d, 1, 2, 4 and 8 wk and the inorganic P determined (Olsen & Sommers, 1982).

Experiment 6: effect of ‘bioassay soil extracts’

To test their effect on growth of A. arenaria (population Voorne) in nonsterilized V-2 sand, bioassay soil extracts were added weekly during the first 4 wk. Control plants received deionized water only. Plants were harvested after 8 wk with 10 replicates per treatment. Briefly, the preparation of extracts was as follows. Marram grass plants (4 per pot) were grown in the greenhouse for 2–4 wk in 1.5-l pots filled with sterilized or nonsterilized sand at 10% moisture (w/w). Each week, nine pots of both sterilized and nonsterilized sand were extracted, three each of 2-wk-, 3-wk-, and 4-wk-old plants. On average, the fr. wts of these plants were 7%, 11% and 19% higher in sterilized sand, respectively. The whole pot content was transferred to a 3-l container, 250–300 ml of deionized water added, and the slurry gently shaken by hand several times during 1 h. Next, in two successive portions, the sand was extracted in a Buchner funnel under vacuum using Whatman 42 paper. At an average extraction efficiency of 50–60%, the nine pots delivered 2 l of soil extract. One half of the extract was filter-sterilized through a 0.2-µm membrane filter (Schleicher & Schuell ME24; Dassel, Germany). Thus, four types of extract were obtained from two series of planted pots. All extracts were prepared within 2 d and stored at 4°C. Subsequently, the extracts were gradually added (at a total of 100 ml pot−1 wk−1) in portions to the respective series on d 2, 4 and 6 after starting their preparation. The extracts were dispensed directly onto the soil surface, as well as through two pipette tips positioned at different depths close to the roots.

General harvesting procedures, calculations, and analyses

At harvest, plants were rinsed with deionized water, separated into shoots, roots and (C. arenaria) rhizomes, and their dry mass (70°C, 48 h) determined. Following coarse grinding (< 2 mm) of the roots, any remaining sand was mechanically removed by a sieving procedure in combination with further separation by a seedblower. Possible contamination with shell fragments was estimated via Ca2+ analysis of the finely ground (< 0.2 mm) roots following sand removal. CaCO3 equivalents were calculated for the Ca2+ content associated with concentrations > 250 µmol g−1 d. wt, and root dry mass was corrected accordingly. All other tissues were directly ground (< 0.2 mm) and stored until further analysis.

Relative growth rate (RGR) was calculated for the total bioassay period (t2− t1) using initial dry mass estimates (W1) and dry mass at harvest (W2):

inline image

RGRs were calculated similarly for the successive time intervals of expts 1–3, but using the initial dry mass that followed from estimated dry mass at the start of the bioassay and average growth rate(s) calculated for the preceding interval(s). RGRs are expressed as mg g−1 d−1. The ‘median’ dry mass half-way through a given time interval was obtained by applying the interval RGR (for half the interval duration) to the estimated tissue dry mass at the start of the corresponding interval.

Plant samples dried at 70°C were digested in a sulphuric acid/salicylic acid mixture with hydrogen peroxide. In the digests, total N and P were determined colourimetrically on a Technicon Traacs autoanalyser. Potassium and (root tissues only) calcium were measured by atomic absorption spectrometry. Specific absorption rates (SARs) for N, P and K were calculated according to Williams (1948):

inline image

with root dry mass R1 and R2 at times t1 and t2, respectively, and corresponding plant contents M1 and M2.

Root length was calculated for some root samples of expts 1–3 using a Delta-T Scan analysis (Delta-T Devices Ltd, Cambridge, UK) at a resolution of 120 dots cm−1 and a brightness setting of 172. Each root sample was then dried (70°C, 48 h) and weighed.

The numbers of nematodes in soil were assessed by decanting the pot sand contents. The 9-l containers of expt 4d were decanted completely into a large vessel up to a total solution volume of 100 l. While agitated vigorously by a stream of air, a 10-l subsample was taken and used in the further nematode collection. Nematodes from roots were extracted for 48 h in a mist chamber using Baermann funnels. Nematodes were counted and identified to at least genus level by microscopic examination.

Data analysis

Significance of sterile/control contrasts, per harvest or (expt 5) per incubation period, and of extract/control (deionized water) contrasts (expt 6) was analysed using one-way ANOVA (Statistix, NH Analytical Software, St Paul, MN 55117, USA). Plots of interval growth rates vs tissue biomass for the three time series of marram grass were separately fitted for control and sterilized soils using the Linear Fit Selection options of the Curve Fitter procedure of the Slide Write Plus software package (Advanced Graphics Software, Carlsbad, CA 92008, USA).

Results

Experiments 1 and 2 (A. arenaria)

At 8 wk, sterilization effects were largely confined to the calcareous V-2 sand and essentially absent on the acid T sand, where nonsterile : sterile (NS : S) ratios were not significantly different from 1 (Table 2). Sterilization effects were more pronounced in terms of biomass than when expressed as growth rates. Plant growth was 16–19% (biomass) or 3–4% (RGR) reduced in nonsterile V-2 sand. Final plant yield was relatively high in the T-sand series of the first experiment, where all P had been mixed into the acid sand at pot filling. The lower average NS : S biomass ratio of 0.8 for the T sand in expt 2 failed to reach significance due to a high variation in plant yield, which was twice as large as for the series in V-2 sand (CVs of 34–36% and 16–20%, respectively). The root weight ratio (RWR, i.e. the fraction of total plant biomass allocated to roots) was in the range 0.13–0.20 and showed no consistent effect of sand sterilization. Fig. 1(a)–(c) summarizes the three marram grass time series and the graphs enable a comparison of the two groups of plants at similar tissue mass. The growth stimulus in sterilized sand was obvious, particularly during the early stages. However, plants in control sand clearly appeared to recover from this initial temporary growth lag and then continued growth at a rate equal to plants in γ-treated sand.

Table 2.  Biomass allocation and growth rate of Ammophila arenaria and Carex arenaria after 8 weeks of growth in control or γ-irradiated (sterile) soil (expts 1–3). Values are the means of 15 (expts 1 and 3) or 10 (expt 2) replicate plants
Sand typeSand sterilizationDry mass (g plant−1)RWRRhWRRGR (mg g−1 d−1)NS : S ratio biomassNS : S ratio RGR
  1. RWR = root weight ratio; RhWR = rhizome weight ratio. NS : S ratios were calculated by dividing individual biomass and RGR data of both the control and γ-irradiated series by the corresponding overall average of the latter series. *, **, *** Sterilization contrasts significant at the 0.05, 0.01, and 0.001 probability levels, respectively.

Experiment 1 (A. arenaria, Voorne seedlings)
Voorne, V-22.08**0.1572.70.81**0.96*
 +2.580.1575.81.001.00
Terschelling, T3.880.17***80.10.971.00
 +3.990.2080.21.001.00
Experiment 2 (A. arenaria, Terschelling seedlings)
Voorne, V-22.09*0.15*75.60.84*0.97
 +2.480.1378.11.001.00
Terschelling, T1.700.1472.90.800.94
 +2.120.1677.41.001.00
Experiment 3 (C. arenaria)
Voorne, V-12.44***0.19**0.2780.7***0.64***0.90***
 +3.840.210.2889.81.001.00
Figure 1.

Mean relative growth rate (RGR, per time interval) of Ammophila arenaria (a–c; expts 1,2) and Carex arenaria (d–g; expt 3), growing in control (closed circles) and γ-irradiated (sterile; open circles) dune sand, as a function of ‘median’ plant (tissue) size during each time interval. RGR was based on the dry mass of whole plants (a,d), shoots (b,e), roots (c,f) and rhizomes (g). Lines in a–c were fitted using graphics software and r2 values of fitted regressions are (n = 12; control and γ-treated sand, respectively): 0.69, 0.92 (whole plant, P < 0.001); 0.92, 0.94 (shoot, P < 0.001); and 0.69 (P < 0.001), 0.59 (P > 0.05) (root). Note differences in axes scales for the two species.

Experiment 3 (C. arenaria)

After 8 wk, growth of Carex arenaria was significantly reduced in control V-1 sand (Table 2): decreases were on average 36% (biomass) and 10% (RGR). Differences in RWR were small but nevertheless significantly lower (P < 0.001) in control sand: 0.19 vs 0.21. In contrast, the proportion of biomass allocated to the rhizomes (RhWR) was unaffected by sand sterilization. Patterns similar to those of marram grass were found when plotting growth rate against the corresponding tissue dry mass (Fig. 1d–g). An obvious stimulatory effect in sterilized sand was again observed, especially in the shoot, although this was most evident at a later stage than observed for marram grass. Growth rates were relatively high in comparison with marram grass. Sand sterilization had no effect on the high rhizome growth rate.

Specific root length (SRL)

By 8 wk, the SRL of marram grass was mostly within the range 35–55 m g−1 root dry mass, with generally no significant difference for sterilized vs control sand (not shown). At earlier growth stages, values tended to be comparatively higher, especially for plants growing in γ-treated sand. The SRL of C. arenaria was largely within the range 75–90 m g−1 for the entire growth period. No significant differences were found in SRL between Carex plants grown in control vs sterilized sand.

Experiment 4: effect of plant size (A. arenaria)

The positive effect of γ-irradiation on plant biomass and RGR remained when using larger plants (and pots) (Table 3). However, although differences became larger in absolute terms, statistical significance usually decreased. Accordingly, the relative difference between control and sterilized sand (and thus the NS : S ratio) was largest when using the smallest plant size: 0.51 (biomass) and 0.89 (RGR). The RWR was either not affected by sand sterilization or was significantly higher for plants growing in sterilized sand (P < 0.001 and P < 0.05 for bioassays b and c, respectively). The largest difference in RWR between control and sterilized sand coincided with the largest initial RWR (bioassay b).

Table 3.  Biomass allocation, growth rate, and P acquisition in Ammophila arenaria after 8 weeks of growth in control or γ-irradiated (sterile) V-2 sand at different initial plant size (expt 4)
Initial plant dry mass (root weight ratio, RWR)Sand sterilizationDry mass (g plant−1)RWRRGR (mg g−1 d−1)NS : S ratioP uptakeShoot P (µmol g−1 d. wt)Shoot P : N (molar ratio)
BiomassRGR(mmol plant−1)(µmol g−1 root d. wt d−1)
  1. Values are the means of 25 (bioassays a,b) or 15 replicate plants (bioassays c,d). Further explanation as in Table 2.

Bioassay a          
2.9 mg (0.12) 0.91***0.16102***0.51***0.89***0.065***50***72***0.039***
 + 1.780.161141.001.000.15968920.048
Bioassay b          
27 mg (0.43) 4.11***0.14*** 89.7***0.74***0.94***0.219***2854***0.035***
 + 5.570.19 95.11.001.000.36929700.054
Bioassay c          
670 mg (0.20)14.3*0.17* 54.6*0.89*0.96*0.556***1343***0.043***
 +16.20.21 56.81.001.000.75114530.053
Bioassay d          
5.30 g (0.19)42.8**0.18 37.0**0.86**0.92**1.926***11***51***0.046***
 +49.70.19 40.11.001.002.76413620.056

Acquisition of N, P and K

Specific absorption rates (SARs) of N, P and K were calculated for expts 1–4. Since differences between control and sterilized sand were most pronounced for the uptake of phosphate, the major emphasis in the following is on SARP. At similar root dry mass, P uptake of both A. arenaria and C. arenaria was clearly favoured in the γ-irradiated calcareous Voorne sands, especially during the early stages (Fig. 2). Maximum uptake rates of P were approximately 45 µmol g−1 d−1 in sterilized sand and 30–57% higher than the maxima reached in control sand. Patterns for N and K were qualitatively similar (not shown). Maximum uptake rates of N and K in γ-treated sand ranged between 870 and 1090 µmol N g−1 d−1 and between 400 and 430 µmol K g−1 d−1, which was 30–39% (N) and 24–44% (K) higher than the maxima in control sand. In contrast to the calcareous sand types, uptake rates by marram grass in the acid T sand showed no consistent differences between control and sterilized sand (P) or were slightly less in the γ-treated sand (N, K) (not shown). In expt 4, overall rates of P uptake by marram grass in sterilized V-2 sand were enhanced independently of plant size, albeit only significant (P < 0.001) for bioassays a and d (Table 3). Generally, shoot P concentrations of both A. arenaria and C. arenaria showed sharp decreases with plant age and were, for marram grass, always higher in T than in V-2 sand (not shown). Phosphate concentrations and P : N ratios in shoots were also usually significantly higher for plants of both species grown in sterilized sand (shown for expt 4 in Table 3).

Figure 2.

Mean specific absorption rate (SAR, per time interval) of P, for plants growing in control (closed circles) or γ-irradiated (sterile; open circles) calcareous dune soils, as a function of ‘median’ root dry mass during each time interval. (a) Ammophila arenaria (population Voorne) in Voorne (V-2) sand, expt 1. (b) A. arenaria (population Terschelling) in Voorne (V-2) sand, expt 2. (c) Carex arenaria in Voorne (V-1) sand, expt 3.

Experiment 5: soil-P availability

In the absence of plants, availability of both the indigenous and added inorganic P was significantly enhanced in γ-treated sand throughout the 56-d incubation period (generally P < 0.001; Fig. 3). Without added C, overall increases of Olsen P in sterilized vs control sand varied between 0.6 µg g−1 (48%) and 0.8 µg g−1 (19%) for the zero and highest P levels, respectively. With addition of C, the corresponding P increases were 0.5 µg g−1 (60%) and 1.2 µg g−1 (55%).

Figure 3.

NaHCO3-extractable inorganic P in control (closed circles) or γ-irradiated (sterile; open circles) Voorne (V-1) sand after 1–56 d of incubation at 20°C/8% moisture and for different levels of P supply (expt 5). P additions (µg g−1): 0 (a,d), 5 (b,e) or 10 (c,f). No C added (a–c) or 1 mg C per g sand added (d–f). Each data point is the mean of three replicates and error bars (SE) are within the size of the symbols. The P increase in irradiated sand was always highly significant (P < 0.001) except where indicated (*P < 0.05; **P < 0.01).

Experiment 6: effect of soil extracts (A. arenaria)

Growth of A. arenaria in nonsterile V-2 sand showed two very different responses to the addition of ‘bioassay extracts’. While average shoot d. wt was not affected, remaining in the range 3.42–4.02 g, root biomass (Fig. 4) was significantly (P < 0.05) enhanced by nonfilter-sterile extracts of planted sterilized sand. Interestingly, addition of extracts from planted nonsterile soil, which had not been filter sterilized, tended to result in a further decline in root yield (P < 0.10) in the nonsterile soil. The 0.2µ-filtration eliminated the effects completely.

Figure 4.

Effects of ‘bioassay soil extracts’ on root d. wt of Ammophila arenaria after 8 wk of growth in nonsterilized Voorne (V-2) sand (expt 6). Extracts were added during the first 4 wk (100 ml pot−1 wk−1): control = deionized water, PSS = planted sterilized sand, PCS = planted control sand, 0.2µ= filter sterile. Values are means of 10 replicates (bars show standard error). Significant differences with the control treatment are indicated: *P < 0.05.

Nematodes in control pots

As expected, no nematodes were found in the sterilized sand series. Nematode counts in nonsterile pots yielded highly variable results, both within and among experimental series (not shown). Nonplant-feeding (and mainly bacterial-feeding) nematodes reached their highest numbers in expt 6 with the addition of bioassay extracts, followed by the T-sand series of expt 2, yielding up to > 100 000 and 50 000 per pot, respectively. Hyphal-feeding taxa of the Aphelenchoididae and the genus Aphelenchus were relatively abundant in expts 3, 4 (bioassays a and b) and 6 (all extract treatments): 500–2500 per pot. In only a few instances did the total number of plant-feeding nematodes exceed 100 per pot (soil + roots). Major species and taxa found were Pratylenchus sp. (20–30 per g root dry mass; expt 4), Tylenchorhynchusmicrophasmis (12 000 per pot, T-sand series of expt 2; 150 per pot, expt 3) and Tylenchidae (300–600 per pot; T-sand series of expt 2 and expt 6).

Discussion

Bioassays using sterilization or biocide-amendment of soils are frequently applied to demonstrate the presence of negative biotic factors in soil, where a positive sterilization effect (or NS : S ratio < 1) suggests the elimination of soil-borne pathogenic effects. Regarding the role of soil-borne pathogens in steering plant succession in dune ecosystems, biomass NS : S ratios have been critical (Van der Putten et al., 1993). Biomass NS : S ratios of A. arenaria and C. arenaria in the present study were similar or (much) higher than those reported in or inferred from the literature for essentially the same sand substrates (Van der Putten et al., 1993; D’Hertefeldt & Van der Putten, 1998). Such inconsistent observations may suggest some dependence of the NS : S ratio on initial plant size and nutrient regime. In many previous bioassays, the addition of nutrients has been comparatively less and provided (in the end) at a linear rather than increasing rate, which may easily cause a relative difference. Also, at very small seedling size, establishment (rather than growth) becomes part of the bioassay. Attributing the ‘sterilization effect’ to plant pathogens is only valid in the absence of other sterilization artifacts that may affect plant growth. The key question thus remains whether the early growth stimuli (Fig. 1) should be related to some (negative) biotic factor in the control soil, to some (positive) (a)biotic characteristic in the sterilized soil, or perhaps to both.

As compared with control soil, there is a whole array of changes in γ-treated soil (also during its microbial recolonization) involving both chemical and microbiological factors (Cawse, 1975; Powlson & Jenkinson, 1976) and irradiation-induced growth increases are common (Cawse, 1975; Jakobsen & Andersen, 1982). At a given rate of nutrient supply, differential reactions in sterilized vs control soils may very well have different nutritional consequences for plant growth, especially when the overall supply has been suboptimal. At suboptimum supply of N or P, the plant’s growth rate is very tightly controlled over quite some range by the effective addition rate of these nutrients (Ingestad, 1987; Ingestad & Ågren, 1988). Plant growth may even increase over a range of nutrient concentrations in the soil solution, which would be predicted to be more than adequate (Robinson et al., 1991). The present study has confirmed an improved nutrient status (higher specific absorption rates, higher shoot nutrient concentrations) for the larger root and shoot compartments in sterilized sand. Once enhanced in biomass, photosynthesis may be facilitated through increased leaf area and higher shoot N and P concentrations (Ozanne, 1980; Lambers & Poorter, 1992), while the root will acquire nutrients more easily. The higher SAR, at equal biomass, for roots in sterilized sand (Fig. 2) strongly suggests a higher intrinsic availability of essential nutrients in this sand, which may be of particular value for the poorly mobile phosphate. Indeed, our results convincingly support the view that the plant’s increased acquisition of P in irradiated sand is at least in part due to its higher availability in the soil (Fig. 3). Apparently, the applied nutrient regime cannot compensate for the relatively impaired nutrient acquisition and retarded growth in control sand. For shoots of control plants in calcareous Voorne sand, both the final P concentration and P : N molar ratio suggested a relative P shortage (Table 3). Although dependent in part on plant age, tissue P (µmol g−1 d. wt) in the ranges of 95–160, 50–90 and < 35 are considered to be optimal, critical and severely limiting, respectively (Kamprath & Watson, 1980; Ozanne, 1980; Troelstra & Brouwer, 1992; Marschner, 1995). Indicative optimum P : N molar ratios for agricultural crops and tree species without luxury uptake are in the range 0.05–0.08 (Dijkshoorn & Lampe, 1980; Ingestad, 1987).

Besides the occurrence of a direct nutrient flush, the temporary elimination of the microbial biomass, and thus the temporary absence of any immobilization, may also cause a temporary increase in the availability of nutrients. Another common phenomenon of irradiated soil is the relative increase of ammonium in the mineral N pool, which also may contribute to enhanced plant growth as compared with exclusive nitrate supply (Cox & Reisenauer, 1973; Troelstra et al., 1995). Therefore, under a nitrate-N nutrient regime, the (temporary) lack of especially nitrification in sterilized soil may indeed stimulate plant growth, contrary to suggestions by Olff et al. (2000). Such an enhancement may have been less effective in the acid T sand due to its intrinsically high relative NH4+ concentration (Table 1) and relatively low nitrification potential at a relatively high N mineralization rate (not shown), as compared with Voorne sand.

Our and other bioassay results are consistent with those of Jakobsen & Andersen (1982) for N-fertilized barley in a sand/soil mixture (with no indication of any specific pathosystem) and up to 40 kGy irradiation. For nonmycorrhiza-inoculated plants, their data indicated similar biomass NS : S ratios (as low as 0.5 – for periods up to 13 wk) and qualitatively similar patterns of shoot growth as presented here. Plants in irradiated soil showed a highly improved P status and increases in barley growth were attributed to elevated concentrations of both plant-available soil P and N. In the various bioassays of expt 4, the consistent positive sterilization effect correlated well with an effective increase in P acquisition in the sterile sand in the order of 1.5–3.0 µg g−1. Indeed, P increments of 0.5–1.5 µg g−1 are no exception (Fig. 3) and plant P uptake usually increases more than would be expected from soil P analyses (Jakobsen & Andersen, 1982).

So far, bioassays with dune plants have not identified specific pathogens or pathogen complexes (nematodes and/or fungi). Several greenhouse experiments with sterilized dune soils have indicated that ectoparasitic nematode species may affect plant growth negatively. However, growth reduction comparable with that in nonsterile soil occurs only at unnaturally high numbers (e.g. Maas et al., 1983; De Rooij-Van der Goes, 1995). Although a more damaging role might be assigned to the endoparasites Pratylenchus, Heterodera and Meloidogyne (Van der Putten & Van der Stoel, 1998; Van der Putten, 2000), they are not always present in the performed bioassays. In the present study, Pratylenchus sp. was essentially the only endoparasitic nematode found and it was present in only some of the experiments at low densities. Therefore, any direct short-term effect of nematodes within the duration of a bioassay must be considered very unlikely (Van der Putten & Troelstra, 1990). Assuming a biotic origin to positive sterilization effects, it still remains open to debate whether the phenomenon should be attributed to general weak ‘minor pathogens’ or to interacting complexes of potentially more destructive, aggressive parasites that are individually somehow less active (Van der Putten, 2000).

Although difficult to fully interpret, the results of expt 6 with only control sand indicated that factors other than nutrient supply may also be involved. Apparently, both the nonfilter-sterile extracts of planted sterilized or planted control sand contained some factor of biotic (bacterial?) origin, which produced quite contrasting effects. This is a first direct suggestion that the (selective) recolonization of the soil microflora following soil sterilization may produce some plant (root) growth promoting effects. If indeed operative in the sterilized sand, it may facilitate the plant’s nutrient capture.

In summary, evidence has been presented that the short-term stimulus of plant growth in γ-irradiated soil may be associated with an improved nutrient acquisition (especially P) due to an increased nutrient availability, enhanced root growth, or both. Apparently, this may even occur when nutrients are applied in what should be moderate excess. Such an effect thus provides an alternative, nonpathogen-related explanation for the positive radiation-induced growth responses. Nutrient-related differences in plant growth may especially occur in those bioassays where initial weekly nutrient supplies are less than the instantaneously released nutrients upon γ-irradiation. Future research must recognize these differential patterns of nutrient availability in sterilized and control sand. In addition, a fair comparison between control and irradiated soil might demand (1) that NH4+ be made part of the N supply, and (2) that control soils be given a proportionally increased amount of nutrients. ‘Real’ biotic effects can only be measured after these issues have been properly addressed. The ecological speculations involving the role of soil-borne biological factors should be based upon fully validated bioassays that account for potential nutritional side-effects induced by soil sterilization.

Acknowledgements

We thank Cindy van Santen for her excellent technical assistance with expts 3 and 4; and Wietse de Boer, George Kowalchuk, Wim van der Putten, Hans van Veen and Jan Woldendorp for valuable comments on the manuscript. Publication no. 2748 of the Netherlands Institute of Ecology.

Ancillary