Soil close to roots generally has chemical, physical and biological properties which are significantly different from those of soil located some distance away (Jenny & Grossenbacher 1963; Hawes & Pueppke 1986; Young 1995). The root–soil interface is defined as soil near to or adhered to the root surface to some small distance into the soil matrix. This distance may vary between <1 mm and c. 10 mm. Working definitions include rhizosphere, where ‘zones of influence’ are inferred, and rhizosheath, when soil adhered to the root is discussed. Most work carried out at the root–soil interface has concentrated on biological or chemical processes, due both to the fact that the relevant techniques required to examine these processes have been more advanced than the physical techniques, and also because the farmer is generally offered either biological or chemical solutions to his everyday problems, as these are readily accessible, easy to use and cheap. The main manipulation of soil physical conditions occurs during cultivations, and the addition or removal of water from the soil profile. Intensive cultivations are a regular occurrence in many farming systems, despite the potential drawbacks: damage of the soil structure, leading to reduced crop yields and an increased risk of erosion.

The main aim of this review is not to cover all the complex issues related to the root–soil interface. Instead, it concentrates on the biophysical processes which, compared with conventional plant physiological and soil microbiological research, have attracted relatively limited attention (e.g. see Waisel et al. 1996). The review examines small-scale (μm-mm) interactions and, where possible, links their impact to the larger scale. Three interacting areas are investigated: the physical structure of the soil and root growth, water dynamics and microbial dynamics.

Great bustard (Otis tarda) populations in Spain are located mostly in cultivated areas during winter. The birds concentrate in large flocks and feed on cultivated plants. This paper reports two types of bustard-caused damage which occur in cereal, lucerne and vetch crops in west-central Spain, as a result of grazing and other bird activities such as display. Field work was carried out between 1994 and 1995. Damage to cereal crops was caused during germination and early seedling growth or to young plants, although plants were able to recover from this by regrowth and the overall yield was not reduced. Vetch crops were damaged in localized areas, but these recovered and were useful for forage enrichment or as grain. However, the damage to lucerne plots, although local, was extensive, resulting in yield losses of 23–31% with respect to the total production.

Diverse cultivars of winter wheat (Triticum aestivum L.) were grown in the field in 1993/94 and 1994/95 at Reading UK in temperature gradient tunnels at normal atmospheric (c. 370) or elevated CO2 concentration (c. 700 μmol CO2 mol−1 air). In 1993/94, grain yield of cv. Avalon was insensitive to mean temperature (between 8·8 and 10·9°C), while elevated CO2 increased yield by 1·3 t ha−1 (12·6%). In all other cultivars, warming reduced grain yield and CO2 increased grain yield. In 1993/94, in cvs Galahad and Mercia the effects of CO2 and temperature on yield were additive. However, for cv. Hereward in both years and for cv. Soissons in 1994/95, there were negative interactions between the effects of CO2 and temperature on yield: the maximum benefit of doubling CO2 to grain yield, 4·5 and 2·7 t ha−1 (65 and 29%) respectively, occurred at cooler temperatures; there was no benefit from doubling CO2 (i.e. 0%) once the temperature had increased above the seasonal mean by 2·2–2·6°C in cv. Hereward and by 1·3°C in cv. Soissons. The beneficial effect of doubling CO2 on grain yield in cvs Galahad, Hereward, Mercia and Soissons was negated by an increase in mean seasonal temperature of only 0·7–2·0°C. Warming decreased root dry mass at anthesis in 1994/95 while it increased at elevated CO2 (49 and 186%, coolest and warmest regime, respectively). Carbon partitioned to roots declined progressively with warming, while at elevated CO2 there was an average of 56% increase in allocation to roots. The relative impacts of both CO2 and temperature were greater on root dry mass than on either grain yield or total above-ground biomass, while the effects on grain and biomass yield varied considerably between cultivars, suggesting that the impact of rising CO2 and temperature are likely to be dependent on cultivar.

The response of cultivars to applied nitrogen was examined in 11 seasons, 1982–92, in two experiments per year, normally testing seven cultivars at seven rates of fertilizer nitrogen. In all, 27 cultivars were tested in 22 experiments throughout Nottinghamshire, Lincolnshire, Northamptonshire and Suffolk. Cultivars ranged in their date of introduction from Maris Huntsman (1969) to Hereward (1988). For each cultivar in each experiment, the economic optimum yield (Yopt), the amount of fertilizer N needed to produce it (Nopt), the grain %N at Nopt, the offtake of N in the grain at nil N (Noff(N0)) and Nopt (Noff(opt)) and the estimated recovery of fertilizer in the grain at Nopt (AFRopt) were estimated by fitting linear plus exponential curves to data for grain yield and two-straight-line models to data for grain N offtake. From cross-site analysis, normalized cultivar means were calculated for each variate. Over the 20-year period relating to the cultivars in the trial, the contribution of new genotypes to grain yield improvement was 1·92 t/ha, Yopt increasing by 96 kg/ha per year. There was no change in grain %N at Nopt. The effect of changes through breeding from 1969 to 1988 was to increase Noff(opt) by 42 kg/ha (2·1 kg/ha per year), that was associated with a decrease in Noff(N0) (equivalent of soil N offtake) of 15 kg/ha (0·77 kg/ha per year). Part of the increased requirement for fertilizer N was fulfilled by an increase in AFRopt of 18% over the 20-year period. The net effect was for Nopt itself to increase by 56 kg/ha (2·8 kg/ha per year). Since survey evidence indicates no general increase in N use on wheat by farmers since the mid-1980s, it appears that current fertilizer use by farmers may be underestimating the requirement for N now. Alternatively in previous years N requirements may have been overestimated. The change in N available for loss to the environment, from the balance of grain Noff(opt) and Nopt, was from 11 kg N/ha in 1969 compared to 25 kg N/ha in 1988. It seems possible that the potential increase in nitrate levels in groundwater associated with plant type may not have been realised because farmers have conserved the amount of N they use.

The relationship between initial population density of Globodera pallida and yield was examined for five genotypes of potato at four different experimental sites; two in Scotland and two in England. The experiments at each site were conducted in two stages. The experimental area was first manipulated using resistant and susceptible potato cultivars to give plots with different nematode population densities to act as the initial populations (Pi) for the second stage. The main experiment was conducted using a range of host genotypes that differed in their tolerance and susceptibility. The utility of an inverse linear model relating yield to initial population density was confirmed as were differences in cultivar tolerance. Site differences in yield loss were also observed. It was found that the model could be modified to include the partitioning of genotype and site effects. The implications of this in terms of developing further models to predict yield loss and nematode population dynamics are discussed.

Between 1989 and 1993, 17 experiments tested the effect of cover crop species, sowing date and destruction date on cover crop dry matter (DM) yield, N uptake and on soil mineral nitrogen (SMN) content. All the experiments were carried out in Suffolk, Norfolk, Lincolnshire and Yorkshire on sandy-loam textured soils after crops of cereals or oilseed rape had been harvested. The largest DM yields were obtained with early sowings and averaged 1·6 t/ha. Cover crop N uptake was less dependent upon sowing date and averaged 35 kg N/ha. The average reduction in SMN was from 46 to 32 kg N/ha. Differences between cover crop species were small when compared with season/site variations.

Cereal cover crop DM yields were closely related to the thermal time accumulated from the first significant rainfall after sowing, whilst the yields of non-cereal cover crops were more affected by the moisture content of the soil at sowing. The amount of SMN in the soil at sowing had little or no effect on cover crop yield. The yields of cereal cover crops were much more predictable than those of non-cereal cover crops. Water usage by cover crops was estimated to be 20 mm/t DM and large cover crops delayed the onset of leaching and reduced the amount of water leached. However, even in dry autumns and winters, soils are likely to reach field capacity before the following beet crop is sown. Due to their large C[ratio ]N ratio (20[ratio ]1) little N would be mineralized after cover crop destruction. Cover crops comprising volunteer cereals and weeds often performed as well as the other cover crops and in most cases will be the most cost-effective cover crops.

Between 1989 and 1993, 17 experiments tested the effects of autumn-sown cover crops on the yield, processing quality and N nutrition of subsequent sugarbeet crops. Cover crops had no effect on sugarbeet plant population density or pesticide requirement. In nitrogen response experiments, the mean beet yield at the economic optimum was 83 t/ha. The mean N fertilizer requirement was 96 kg N/ha and the N uptake at maximum yield averaged 180 kg N/ha. Cover crops had no effect on yield, fertilizer requirement or N uptake. In addition, cover crops generally had no effect on the efficiency of N fertilizer use, the mineralization of N from the soil organic matter nor the amount of soil mineral nitrogen at sowing or at harvest of the beet crop. Processing quality was also not affected by cover crops. The cost of growing a cover crop ranged from 0 to 50 £/ha. Since these costs cannot be offset against increases in yields or reduced fertilizer application rates, cover crops need to be low cost, i.e. cheap seed and minimal cultivation. Cover crops using volunteer cereals and weeds or farm-saved grain that are established with a single stubble-cultivation should fulfil these criteria.

Solid pig manure (240 g kg1 DM) and solid cattle manure (150-180 g kg1 DM) were stored in an open storage facility during spring-summer and autumn conditions for periods of 9-14 weeks during 1994 and 1995. Concentrations of C, N, P and K were determined prior to and after storage, corrected for dry matter losses and distance from the surface. Temperature and, in experiments with pig manure, gas phase composition inside the manure heap were monitored during storage. Nitrogen losses as ammonia volatilization, nitrous oxide emission and leaching were measured, while total denitrification was estimated from mass balance calculations. For both cattle and pig manure there was little difference between seasons with respect to the pattern of decomposition, as reflected in temperature dynamics and C/N turnover. In contrast, there was a distinct difference between manure types. Pig manure was characterized by maximum temperatures of 60-70°C, although the concentrations of oxygen and methane clearly demonstrated that anaerobic conditions dominated the interior parts of the heap for several weeks. Losses of C and N from pig manure both amounted to c. 50%. In contrast, the temperature of cattle manure remained close to the air temperature throughout the storage period and cattle manure had lower, not significant losses of C and N. Leaching losses of N constituted 1-4% with both manure types. Ammonia volatilization from cattle manure constituted 4-5% of total N, and from pig manure 23-24%. In pig manure a similar amount of N (23-33%) could not be accounted for after storage, a loss that was attributed to denitrification. Nitrous oxide emissions amounted to <2% of estimated denitrification losses.

The effects of dietary variations on microbial N turnover and recycling related to protozoal activities in the rumen were examined using a previously described model of rumen fermentation processes. Simulations were performed with diets containing variable proportions of concentrate (between 0 and 1 kg/kg diet dry matter) at a rate of daily dry matter intake (DMI) by cattle of between 5·3 and 21·0 kg/day. The roughages examined were hay, fresh grass and maize silage, and the concentrate supplements included molasses, barley, maize grain, untreated and formaldehyde-treated soyabean meal. The simulated fractional turnover rate of microbial N varied between 0·416 and 1·556/day, and simulated microbial N recycling ranged from 34·9 to 76·2% of gross microbial N synthesis. High turnover rates and recycling were observed particularly for diets containing maize silage as roughage and molasses and maize grain as concentrate supplements. Increased DMI generally resulted in reduced turnover rates with high-roughage diets, but increased turnover rates with high-concentrate diets. These responses of microbial N turnover and recycling were related to substrate availability and the substrate preferences of protozoa and, consequently, protozoal activities in the rumen. The large impact of recycling on the amount of organic matter required per unit of microbial protein washed out of the rumen was demonstrated. Thus, by quantification of the interactions between microbial populations in the rumen, the model can help to interpret the differences in microbial N recycling observed in the experiments and provide an improved understanding of the efficiency of microbial protein synthesis.

Brahman Shorthorn steers (213±4·6 kg initial weight) were grazed on a pangola (Digitaria eriantha cv. Steudel) pasture interplanted with leucaena (Leucaena leucocephala cv. Cunningham) in rows 1·8–2·7 m apart in the Ord River Irrigation Area (ORIA) in north-west Western Australia. The leucaena comprised a mean of 46% of the diet but ranged from 100% of the diet on day 1 to c. 18% on day 7 of a 7-day grazing period, within the rotational grazing system. It was suggested that the low herbage allowance of leucaena may have restricted the intake of the steers, reducing the quantity and quality of the herbage consumed.

The steers were also supplemented with maize during the dry and wet seasons (1992 and 1992/93). The maize supplement was offered at 0, 0·5, 1·0, 1·5 and 2·0 kg/head per day, over a 168-day period. Maize supplementation at 1·5 kg maize per head per day resulted in an increase in the liveweight gain of the steers grazing leucaena/pangola pastures in the dry season from 0·73±0·05 kg/day (mean±S.E.) for no supplementation to 1·1±0·05 kg/day for 1·5 kg/head per day of maize. Maize supplementation up to 2·0 kg/day did not further increase the liveweight gain. A high rate of substitution of leucaena and pangola intake for the maize supplement occurred.

There was a large seasonal difference in liveweight gain (mean 0·89 v. 0·63 kg/day for the dry and wet seasons respectively; P<0·001) irrespective of the level of maize supplementation. This seasonal difference in liveweight gain was not primarily a function of the herbage on offer, but was most probably a direct function of the effect of the seasonal temperature and humidity influencing the feed intake and growth rate of these Brahman crossbred cattle.

This system produced some of the higher liveweight gains recorded per hectare for a grazing system (1570–2110 kg/ha per year), as a result of the very good individual liveweight gains (250-340 kg/head per year) and high stocking rate (6·25 head/ha).

The effect of concentration of sulphur (S) in a diet on the interrelationship between copper (Cu) and selenium (Se) metabolism was investigated in sheep. The animals received a sheep finishing diet in individual feeding pens for a period of 74 days. The minerals were included in the diets according to a 3×2×2 factorial experimental design. The final mixtures contained three concentrations of Se; 0·35, 0·88 and 1·34 mg Se (Se added as Na2SeO3); two concentrations of Cu; 6·7 and 17·0 mg Cu (Cu added as CuSO4) and two concentrations of S; 2·15 and 3·97 g S (S added as Na2SO4)/kg dry matter (DM). An increase in dietary Cu resulted in a significant (P<0·01) increase in liver Se concentration; 2·30 v. 3·43 mg Se/kg DM for the low v. the high Cu treatments. The addition of S reduced the hepatic Se concentration significantly (P<0·01). This was most pronounced at the high Cu intake, resulting in a significant (P<0·05) three-way interaction between the three minerals, i.e. the hepatic Se concentrations were 2·53, 2·08, 4·07 and 2·79 mg/kg DM for the low Cu–low S, low Cu–high S, high Cu–low S and high Cu–high S treatments, respectively. Dietary S reduced (P<0·01) the concentration of Cu in the liver from 678 mg/kg DM at the low S to 305 mg/kg DM at the high S intakes. Dietary Se did not change the concentration of Cu in the liver significantly. Adding S to the diet caused a reduction (P<0·01) in the Se concentration of rumen bacteria, i.e. from 3·48 mg/kg DM in the low S to 1·96 mg/kg DM in the high S treatments. Selenium concentrations in rumen bacteria and the liver were the most sensitive to changes in Se intake, followed by the heart, muscle, plasma and whole blood. It was concluded that an increase in S intakes would decrease the accumulation of both Cu and Se in the liver of sheep, resulting in a diminishing degree of interaction between Cu and Se.

Experiments were carried out at Uhříněves (Czech Republic) in 1996 and 1997 to evaluate the effects of avoparcin, bacitracin, flavomycin, tylosin and virginiamycin on in vitro fermentation of ovine ruminal contents and rabbit caecal contents. In incubations of the rumen contents supplied with starch, hemicellulose and pectin, all antibiotics at 20 μg/ml decreased production of fermentation gas and changed molar composition of fermentation products in favour of propionate. Flavomycin, tylosin and virginiamycin significantly decreased total production of volatile fatty acids (VFA). In incubations of the rabbit caecal contents, attempts to modify composition of microbial metabolites failed. Bacitracin was the only antibiotic which significantly increased the molar percentage of propionate and decreased acetate. This fermentation shift was accompanied by a decrease in total VFA production. The net production of propionate was thus similar in control and bacitracin-treated cultures. Rumen in vitro fermentation yielded more fermentation gas, methane and propionate than rabbit caecal fermentation of the same substrate. There was no apparent effect of antibiotics tested on methanogens in either microbial system, as the ratios of methane to VFA production were similar in control and treated cultures.