Maintenance and Impacts of an Inoculated mer/luc-Tagged Pseudomonas fluorescens on Microbial Communities in Birch Rhizospheres Developed on Humus and Peat

Antagonistic bacteria represent promising biocontrol agents for improving forest production in seedling nurseries or forest soils. The fate of an introduced mer/luc-tagged antagonistic Pseudomonas fluorescens 31K3 was monitored in the rhizosphere of silver birch (Betula pendula) seedlings grown in microcosms containing forest humus or nursery peat. The inoculated strain (10(8) cfu g(-1) soil) was unable to establish in significant numbers in either soil type and turned nonculturable in humus. Detection in both soils was possible only via luminescence of enrichment cultures 80 days post-inoculation. Despite low P. fluorescens survival, inoculation had a positive effect on seedling growth. Limited impact of inoculation on the indigenous microbial communities was identified following analyses of respiration and denitrification potential, community-level physiological profiles and molecular fingerprinting of fungi and eubacteria, and Pseudomonas community structures. The minor changes observed in the indigenous microbial communities, including mycorrhiza development, were not consistent between humus and peat growth substrates. It was concluded that the rhizosphere-related microbial communities developed in both of these highly organic soil systems are highly buffered against introduction of foreign bacteria.     

Application of bioluminescence-based microbial biosensors to the ecotoxicity assessment of organotins

The toxicity of two common organotin pollutants and their initial breakdown products (tributyltin, dibutyltin, triphenyltin and diphenyltin) were assessed using two different bioluminescent microbial biosensors: Microtox and lux -modified Pseudomonas fluorescens pUCD 607. The organotins were made up as standards, and tested both in double-deionized water and in extracted soil solution, the latter representing a realistic matrix for terrestrial contamination. Microtox was especially sensitive to the organotins, with 50% effective concentration (EC₅₀) values (15 min) for tributyltin as low as 21·9 μg l⁻¹ in pure water, and 0·118 μg l⁻¹ in soil extract. The sensitivity of Microtox was increased by an order of magnitude in soil extract. The Ps. fluorescens was less sensitive, with EC₅₀ values (30 min) of around 800 μg l⁻¹ in pure water. The toxicity to Ps. fluorescens was decreased by around an order of magnitude in soil extract. The two biosensors showed different response patterns, with Microtox being more sensitive to the triorganotins and Ps. fluorescens being more sensitive to the diorganotins.     

Indigenous microflora responses to introduction of cyanogenic strains of Pseudomonas fluorescens into soil.

The effects of cyanogenic Pseudomonas fluorescens strains introduced into soil on the kinetic of colony formation and bacterial community structure were investigated. About 7.8 x 10(8) and 1.2 x 10(9) cfu per g dry soil of TA1 and B2 were added to the soil portions, respectively. The parameters of colony formation by heterotrophic soil bacteria were determined. The bacterial community structure and phenotypic diversity were studied using concept of r/K strategies and echophysiological index, respectively. The physiological state of indigenous heterotrophic bacteria and gram-negative group did not change under the influence of the cyanogenic strains introduced. Phenotypic diversity of the soil bacteria also did not change significantly. However, some short-term shifts in community structure of indigenous heterotrophic bacteria were noticed. This study shows that the introduction of great numbers of cyanogenic P. fluorescens strains could be safely used as potential agents in biological control of soil-born pathogens.     

Plasmid and chromosomally encoded luminescence marker systems for detection of Pseudomonas fluorescens in soil

Luminescent strains of Pseudomonas fluorescens 10586 were constructed in which luciferase production was constitutive by introduction of Vibrio fischeri luxABE genes on the chromosome and on a multicopy plasmid. Light production in liquid batch culture was directly proportional to biomass concentration during exponential growth and enabled detection by luminometry of 1.7 × 10³ and 8.9 × 10⁴ cells/ml for the plasmid and chromosomally marked strains, respectively. Luminescent colonies of both strains were detectable by eye, enabling viable cell enumeration on solid media against a background of non-luminescent strains. Following inoculation into sterile and non-sterile soil lower levels of detection were increased but detection of 8.1–59 × 10³and 2.2–30 × 10³ cells per g of soil was possible for plasmid and chromosomally marked strains. Maximum specific growth rate in liquid culture was unaffected by introduction of lux marker genes on the chromosome, but was reduced in the plasmid marked strain. The chromosomally encoded marker was stable in both liquid culture and in soil, but the plasmid was unstable during continuous subculturing in liquid medium and during growth in soil. The chromosomally encoded luminescence-marker system therefore provides a convenient, non-extractive technique for quantification of genetically modified soil microbial inocula.     

Direct detection of rhizosphere-colonizing Pseudomonas sp. using an Escherichia coli rRNA promoter in a Tn7-lux system

Abstract A promoterless Tn7- lux system conferring bioluminescence was fused with an Escherichia coli rRNA gene promoter and compared with neo - or lac-luxCDABE analogs after introduction in Pseudomonas cells. Fusion of the ribosomal promoter with luxCDABE genes increased the bioluminescence of cells by approx. 100- to 1500-fold over the neo-lux system depending on the growth conditions and bacterial strain. When the cells were grown in suspension culture, light production and growth were strongly dependent on the nutrient composition of the medium. Root-colonizing competence was tested in nonsterile soil by autophotographic detection of bacterial bioluminescence on plant roots. The lower detection limit of the autophotographic method for roots inoculated with Pseudomonas fluorescens 2–79 was 10⁵ cfu g⁻¹ fresh root weight. The new bioluminescence marker did not require addition of supplemental nutrients or the aldehyde substrate for the luciferase enzyme and provides a simple and highly sensitive detection method for long term in situ studies on the microbial ecology of specific bacterial strains.     

Detection of viable, but non-culturable Pseudomonas fluorescens DF57 in soil using a microcolony epifluorescence technique

Abstract Kanamycin-resistant Pseudomonas fluorescens DF57-3 cells (Tn5 modified) inoculated in soil microcosms rapidly lost their culturability, as defined by visible colony formation on Kings B agar supplemented with kanamycin. Thus, after 40 days only 0.02–0.35% of the initial inoculum was culturable. A microcolony epifluorescence technique was developed to determine the viable, but non-culturable subpopulation. A suspension of bacteria from the soil was prepared in salt solution after a sonication procedure and a sample was filtered onto a 0.2 μm Nuclepore filter. The filter was then placed for 3–4 days on the surface of Kings B agar before staining with acridine orange for epifluorescence microscopy. By staining and washing the filters carefully, disruption of microcolonies could be avoided. A majority of the microcolonies resulted from 2–3 cell divisions during the first 2 days of the incubation period, after which the cell divisions stopped. These microcolonies were taken to represent a population of viable, but non-culturable cells and comprised about 20% of the initial inoculum. A similar recovery was obtained when the filters were incubated on the surface of citrate minimal medium or soil extract medium. A few microcolonies showed continued growth on the filters, however, and their number corresponded well with that of visible macrocolonies. Observation by microscopy of a few (2–3) cell divisions (microcolony epifluorescence technique) is proposed for determination of subpopulations of viable, but non-culturable bacteria in soil.     

Method for spiking soil samples with organic compounds

We examined the harmful side effects on indigenous soil microorganisms of two organic solvents, acetone and dichloromethane, that are normally used for spiking of soil with polycyclic aromatic hydrocarbons for experimental purposes. The solvents were applied in two contamination protocols to either the whole soil sample or 25% of the soil volume, which was subsequently mixed with 75% untreated soil. For dichloromethane, we included a third protocol, which involved application to 80% of the soil volume with or without phenanthrene and introduction of Pseudomonas fluorescens VKI171 SJ132 genetically tagged with luxAB::Tn5. For both solvents, application to the whole sample resulted in severe side effects on both indigenous protozoa and bacteria. Application of dichloromethane to the whole soil volume immediately reduced the number of protozoa to below the detection limit. In one of the soils, the protozoan population was able to recover to the initial level within 2 weeks, in terms of numbers of protozoa; protozoan diversity, however, remained low. In soil spiked with dichloromethane with or without phenanthrene, the introduced P. fluorescens VKI171 SJ132 was able to grow to a density 1,000-fold higher than in control soil, probably due mainly to release of predation from indigenous protozoa. In order to minimize solvent effects on indigenous soil microorganisms when spiking native soil samples with compounds having a low water solubility, we propose a common protocol in which the contaminant dissolved in acetone is added to 25% of the soil sample, followed by evaporation of the solvent and mixing with the remaining 75% of the soil sample.     

Plant mediated interactions between Pseudomonas fluorescens, Rhizobium leguminosarum and arbuscular mycorrhizae on pea

Using undisturbed sandy loam soil cores heavily infested with mycorrhizae, the effects of the antibiotic-producing Pseudomonas fluorescens strain F113 and its non-antibiotic derivative Ps. fluorescens F113G22 on nodulation by introduced and indigenous Rhizobium were studied. Furthermore, the effects of the different microbial inocula on the colonization of the pea roots by mycorrhizae were studied. It was found that Ps. fluorescens F113 enhanced nodulation by Rhizobium fourfold, while the nodules produced were much larger and strongly pigmented (pink) compared with those in other treatments. The proportion of roots colonized by arbuscular mycorrhizae was not significantly affected by the different treatments.     

Effect of inoculation of a TOL plasmid containing mycorrhizosphere bacterium on development of Scots pine seedlings, their mycorrhizosphere and the microbial flora in m-toluate-amended soil

The purpose of this study was to evaluate the influence of introduced bacteria containing a contaminant degrading plasmid on the growth and survival of pine seedlings and mycorrhizosphere microbial flora in contaminated soil. The Pseudomonas fluorescens strain OS81, originally isolated from fungal hyphae in contaminated soil, was supplied with the TOL plasmid pWW0::Km (to generate OS81(pWW0::Km)) and inoculated in humus-soil microcosms with and without pine seedlings mycorrhized with Suillus bovinus. After 3 months of regular treatment with m-toluate (mTA) solutions, the introduced catabolic plasmid was found to be disseminated in the indigenous bacterial population of both mycorrhizosphere and soil uncolonized by the fungus. Transconjugants were represented by bacteria of the genera Pseudomonas and Burkholderia and their number correlated positively with the concentration of mTA applied. Indigenous mTA degrading bacteria with low similarity to Burkholderia species were also enriched in microcosms. They were mostly associated with mycorrhizal soil or fungal structures and virtually absent in microcosms without pines. The total number of Tol(+) bacteria was higher in mycorrhizospheric soil compared with bulk soil. Inoculation with P. fluorescens OS81(pWW0::Km) had a positive effect on the development of roots and fungus in contaminated soil. Both inoculation with the P. fluorescens OS81(pWW0::Km) and mTA contamination as well as the presence of mycorrhized pine roots and fungal hyphae had an effect on the microbial community structure of soil as measured by carbon source oxidation patterns. However, the impact of mTA on the microbial community was more prominent. The study indicates that an effect on plant and fungal development can be obtained by manipulating the mycorrhizosphere. Both introduction of the bacterium carrying the degradative plasmid and the plasmid itself are likely to have a positive effect not only on the organisms involved, but also on bioremediation of contaminated soil, a factor that was not directly monitored here.     

The dynamics of Rhizobium leguminosarum biovar trifolii introduced into soil as determined by immunofluorescence and selective plating techniques

Abstract After the introduction of Rhizobium leguminosarum biovar trifolii into a loamy sand and a silt loam, high recovery percentages were determined using quantitative immunofluorescence. Soil type, but not inoculum density between 10⁴ and 10⁸ cells per gram of soil, significantly influenced the recovery percentage of the immunofluorescence technique. Recovery percentages determined using selective plating were independent of either soil type or inoculum density and exceeded those determined by immunofluorescence.
The serological and genetic markers used for detection were stable during 55 days of incubation in phosphate-buffered saline and soil extract solution. After the introduction of R. leguminosarum biovar trifolii into both sterilized soil types, the population increased to 0.5–1×10⁹ cells per gram of soil, but a decline was demonstrated in non-sterile loamy sand and silt loam during incubation of 90 days at 15°C. Starvation of rhizobial cells in the phosphate-buffered saline and soil extract solution, as well as incubation in both soil types, resulted in a significant decrease in mean cell size.     

Effect of genetically modified Pseudomonas fluorescens strains on the uptake of nitrogen by pea from 15N enriched organic residues

The effects of an antibiotic-producing Pseudomonas fluorescens strain (F113) carrying the marker gene cassette lac ZY and a marked, non-producing strain (F113G22) on the uptake of nitrogen from ¹⁵N enriched organic residues incorporated into a sandy soil were investigated in microcosm studies. Strain F113 produces the antibiotic 2,4-diacetylphloroglucinol (DAPG), while its modified derivative strain F113G22 has DAPG production deleted by Tn 5 mutagenesis. Uptake of nitrogen by pea ( Pisum sativum ) was estimated using isotope-ratio mass spectrometry. In addition, plant growth and microbial activity in soil were monitored. Both strains F113 and F113G22 enhanced the uptake of nitrogen from mineralized organic residues, even though the antibiotic producing strain F113 significantly reduced microbial activity in soil. It is suggested that the effect on nitrogen uptake was due to increased mineralization of organic residues by the introduced organisms, making greater quantities of inorganic nitrogen available for plant uptake. Unlike studies assessing impact in terms of perturbation to indigenous microbial communities, this study provides direct evidence of a change in ecosystem function as a result of the introduction of strains of a genetically marked bacterium, irrespective of whether its natural antibiotic-producing capacity has been genetically deleted.     

Effect of cell density and attachment on resuscitation in soil of starved Pseudomonas fluorescens MON787

The effect of cell density and attachment on starvation survival and recovery was determined using luminometry to measure activity of a lux -marked strain of Pseudomonas fluorescens MON787. Bioluminescence was found to be a sensitive indicator of in situ activity of P. fluorescens MON787 in soil. The activity of a bacterial inoculum could be monitored during growth in soil, and was found to correlate with an increase in cell numbers. Luminescence could detect decreasing activity of P. fluorescens during starvation in soil, and recovery of activity and cell numbers following exposure to starvation and matric potential stress. The effect of localised cell density and attachment in soil on recovery from lag phase after nutrient addition was investigated and compared to recovery of starved liquid cultures. Nutrient addition to starved P. fluorescens in soil or liquid medium resulted in an immediate recovery of activity, followed by a second increase in luminescence after 5 h. Cells exposed to both starvation and matric potential stress in soil did not show a detectable immediate increase of activity, but required a 5-h lag phase before recovery of both activity and cell growth. The lag phase values were not significantly different over a range of localised cell densities. This suggests that cell density of P. fluorescens in the range tested is not a factor which affects recovery of soil bacteria from starvation.     

Effect of microbial inoculants on the indigenous actinobacterial endophyte population in the roots of wheat as determined by terminal restriction fragment length polymorphism

The effect of single actinobacterial endophyte seed inoculants and a mixed microbial soil inoculant on the indigenous endophytic actinobacterial population in wheat roots was investigated by using the molecular technique terminal restriction fragment length polymorphism (T-RFLP). Wheat was cultivated either from seeds coated with the spores of single pure actinobacterial endophytes of Microbispora sp. strain EN2, Streptomyces sp. strain EN27, and Nocardioides albus EN46 or from untreated seeds sown in soil with and without a commercial mixed microbial soil inoculant. The endophytic actinobacterial population within the roots of 6-week-old wheat plants was assessed by T-RFLP. Colonization of the wheat roots by the inoculated actinobacterial endophytes was detected by T-RFLP, as were 28 to 42 indigenous actinobacterial genera present in the inoculated and uninoculated plants. The presence of the commercial mixed inoculant in the soil reduced the endophytic actinobacterial diversity from 40 genera to 21 genera and reduced the detectable root colonization by approximately half. The results indicate that the addition of a nonadapted microbial inoculum to the soil disrupted the natural actinobacterial endophyte population, reducing diversity and colonization levels. This was in contrast to the addition of a single actinobacterial endophyte to the wheat plant, where the increase in colonization level could be confirmed even though the indigenous endophyte population was not adversely affected.     

Survival of, and induced stress resistance in, carbon-starved Pseudomonas fluorescens cells residing in soil.

We investigated the survival, cell length, and development of general stress resistance in populations of Pseudomonas fluorescens R2f and its rifampin-resistant mutant, R2f Rpr, following exposure to carbon starvation conditions in liquid cultures and residence in two different soils, Flevo silt loam (FSL) and Ede loamy sand (ELS). In much the same way as was recently shown for P. putida KT2442, carbon-starved P. fluorescens R2f populations revealed enhanced resistance to otherwise lethal treatments, such as exposure to ethanol, high temperature, osmotic tension, and oxidative stress. A large population of nonculturable P. fluorescens R2f Rpr cells arose shortly after their introduction into ELS soil, whereas the formation of nonculturable cells was not observed in FSL soil. Also, the inoculant cell (based on immunofluorescence) and CFU counts decreased faster in ELS soil than in FSL soil. Introduction of carbon-starved instead of exponential-growth-phase R2f Rpr cells into ELS soil did not affect bacterial survival. The inoculant cell length decreased in soil, and no large differences in cell length in the two soil types were observed. Addition of glucose to ELS soil resulted in a stable cell length of R2f Rpr cells, whereas carbon-starved cells introduced into ELS soil remained small. Exponentially growing R2f Rpr cells developed enhanced resistance to ethanol, high temperature, osmotic tension, and oxidative stress within 1 day in both soils, whereas cells introduced into ELS soil amended with glucose showed decreased resistance. Cells that were carbon starved prior to introduction into ELS soil showed unchanged stress resistance levels upon residence in soil.     

Maintenance of soil functioning following erosion of microbial diversity

The paradigm that soil microbial communities, being very diverse, have high functional redundancy levels, so that erosion of microbial diversity is less important for ecosystem functioning than erosion of plant or animal diversity, is often taken for granted. However, this has only been demonstrated for decomposition/respiration functions, performed by a large proportion of the total microbial community, but not for specialized microbial groups. Here, we determined the impact of a decrease in soil microbial diversity on soil ecosystem processes using a removal approach, in which less abundant species were removed preferentially. This was achieved by inoculation of sterile soil microcosms with serial dilutions of a suspension obtained from the same non-sterile soil and subsequent incubation, to enable recovery of community size. The sensitivity to diversity erosion was evaluated for three microbial functional groups with known contrasting taxonomic diversities (ammonia oxidizers < denitrifiers < heterotrophs). Diversity erosion within each functional group was characterized using molecular fingerprinting techniques: ribosomal intergenic spacer analysis (RISA) for the eubacterial community, denaturing gradient gel electrophoresis (DGGE) analysis of nirK genes for denitrifiers, and DGGE analysis of 16S rRNA genes for betaproteobacterial ammonia oxidizers. In addition, we simulated the impact of the removal approach by dilution on the number of soil bacterial species remaining in the inoculum using values of abundance distribution of bacterial species reported in the literature. The reduction of the diversity of the functional groups observed from genetic fingerprints did not impair the associated functioning of these groups, i.e. carbon mineralization, denitrification and nitrification. This was remarkable, because the amplitude of diversity erosion generated by the dilution approach was huge (level of bacterial species loss was estimated to be around 99.99% for the highest dilution). Our results demonstrate that the vast diversity of the soil microbiota makes soil ecosystem functioning largely insensitive to biodiversity erosion even for functions performed by specialized groups.     

Influence of an arbuscular mycorrhizal fungus on Pseudomonas fluorescens DF57 in rhizosphere and hyphosphere soil

The influence of Glomus intraradices (BEG87) on Pseudomonas fluorescens DF57 in hyphosphere and rhizosphere soil was examined. Cucumis sativus (Aminex, F₁ hybrid) was grown in symbiosis with the arbuscular mycorrhizal fungus G. intraradices in PVC tubes, consisting of a central root compartment and two lateral root-free compartments. Two Tn 5 - lux AB-marked strains of P. fluorescens DF57 were used. Strain DF57-P2, which has an insertion of Tn 5::lux AB in a phosphate starvation-inducible locus, was used as a phosphate starvation reporter. Another lux -tagged strain DF57-40E7, which carries a constitutively expressed lux AB fusion, was used as control for strain DF57-P2 and for measuring the metabolic activity of P. fluorescens DF57. A strain of P. fluorescens DF57, which carries a constitutively expressed gfp gene, was used in studies of attachment between the bacteria and the hyphae. G. intraradices decreased the culturability of P. fluorescens DF57 significantly, both in rhizosphere and hyphosphere soil, whereas the total number of P. fluorescens DF57 measured by immunofluorescence microscopy was decreased in hyphosphere soil only. G. intraradices did not induce a phosphorus starvation response in P. fluorescens DF57, and the metabolic activity of the bacteria was not affected by the fungus after 48 h. P. fluorescens DF57 did not attach to G. intraradices hyphae and was not able to use the hyphae as carbon substrate. The negative effect of G. intraradices on culturability and on number of P. fluorescens DF57 in hyphosphere soil is discussed.     

Construction and detection of bioluminescent strains of Bacillus subtilis

Bioluminescence ( lux ) genes from Vibrio fischeri and V. harveyi were introduced into Bacillus subtilis on a plasmid vector and by chromosomal integration. The plasmid-bearing strain was highly luminescent and stable under antibiotic selection, but luminescence was lost in the absence of selection and following sporulation and germination. The chromosomally marked strains emitted less light but were found to be stable without the requirement for antibiotic selection and following sporulation and germination. Individual luminescing colonies of both B. subtilis strains could be detected against a high background of non-bioluminescent indigenous soil microbial colonies on agar plates using a charge-coupled device camera. These bioluminescent Gram-positive strains could be of value in studies concerning the survival and spread of genetically-modified micro-organisms in soil environments.     

Effect of Microbial Inoculants on the Indigenous Actinobacterial Endophyte Population in the Roots of Wheat as Determined by Terminal Restriction Fragment Length Polymorphism

The effect of single actinobacterial endophyte seed inoculants and a mixed microbial soil inoculant on the indigenous endophytic actinobacterial population in wheat roots was investigated by using the molecular technique terminal restriction fragment length polymorphism (T-RFLP). Wheat was cultivated either from seeds coated with the spores of single pure actinobacterial endophytes of Microbispora sp. strain EN2, Streptomyces sp. strain EN27, and Nocardioides albus EN46 or from untreated seeds sown in soil with and without a commercial mixed microbial soil inoculant. The endophytic actinobacterial population within the roots of 6-week-old wheat plants was assessed by T-RFLP. Colonization of the wheat roots by the inoculated actinobacterial endophytes was detected by T-RFLP, as were 28 to 42 indigenous actinobacterial genera present in the inoculated and uninoculated plants. The presence of the commercial mixed inoculant in the soil reduced the endophytic actinobacterial diversity from 40 genera to 21 genera and reduced the detectable root colonization by approximately half. The results indicate that the addition of a nonadapted microbial inoculum to the soil disrupted the natural actinobacterial endophyte population, reducing diversity and colonization levels. This was in contrast to the addition of a single actinobacterial endophyte to the wheat plant, where the increase in colonization level could be confirmed even though the indigenous endophyte population was not adversely affected.     

Distribution of metabolic activity and phosphate starvation response of lux-tagged Pseudomonas fluorescens reporter bacteria in the barley rhizosphere.

The purpose of this study was to determine the metabolic activity of Pseudomonas fluorescens DF57 in the barley rhizosphere and to assess whether sufficient phosphate was available to the bacterium. Hence, two DF57 reporter strains carrying chromosomal luxAB gene fusions were introduced into the rhizosphere. Strain DF57-40E7 expressed luxAB constitutively, making bioluminescence dependent upon the metabolic activity of the cells under defined assay conditions. The DF57-P2 reporter strain responded to phosphate limitation, and the luxAB gene fusion was controlled by a promoter containing regulatory sequences characteristic of members of the phosphate (Pho) regulon. DF57 generally had higher metabolic activity in a gnotobiotic rhizosphere than in the corresponding bulk soil. Within the rhizosphere the distribution of metabolic activity along the root differed between the rhizosphere soil and the rhizoplane, suggesting that growth conditions may differ between these two habitats. The DF57-P2 reporter strain encountered phosphate limitation in a gnotobiotic rhizosphere but not in a natural rhizosphere. This difference in phosphate availability seemed to be due to the indigenous microbial population, as DF57-P2 did not report phosphate limitation when established in the rhizosphere of plants in sterilized soil amended with indigenous microorganisms.     

A determination of protective microhabitats for bacteria introduced into soil

Abstract Survival studies with rhizobia introduced into loamy sand showed that a kaolinite amendment of the soil improved the survival of Rhizobium , and that bentonite had a very strong positive effect on rhizobial survival. The survival level was significantly higher in soil amended with 10% than with 5% bentonite. The amount of water present in the bentonite amended soil had a significant influence on rhizobial survival; in drier soil, survival levels were highest. For the loamy sand, the loamy sand amended with 5 and 10% bentonite or with 10% kaolinite, the number of rhizobial cells surviving on day 57 after introducing 2.5–5.0×10⁷ cells g⁻¹ dry soil could be described using the distribution of pores from three size classes in a mathematical relationship. Pores with necks < 3 μm and between 3 and 6 μm positively affected the survival of introduced rhizobia whereas pores with necks > 6 μm had a negative effect.