Scope 47- Long-term Ecological Research, An International Perspective

6.1 INTRODUCTION

In many parts of the world terrestrial ecosystems are more or less man-managed, and these management inputs interact with climatic change, the physico-chemical properties of the soil, and various other inputs. In the absence of enlightened management, these other inputs can have large effects on soils, plants, and animals; even to the extent of affecting the soil microbial biomass, which acts as a source and sink for nutrients in many natural ecosystems. An example of such an input is acid rain, which has been the subject of much publicity in recent years, and its obviously adverse effect on sensitive ecosystems. The effects of acid rain can be mitigated, however, as the extent of soil acidification depends not only on the amount of acidifying inputs, but also upon the soil properties and the vegetative cover, both of which can be controlled to some extent by correct management. With changing land-use patterns, a changing environment, and increased pressures on the landscape, it is essential to try to assess what the effects of some of these changes may be if irreparable damage to the ecosystems is to be avoided.

Field experiments started in the 1840s to 1860s by Lawes and Gilbert at Rothamsted were designed to find which of the elements present in farmyard manure (FYM) are essential for plant growth, and in what quantity they are essential. The effects of nitrogen, phosphorus, potassium, magnesium, and sodium were tested singly and in various combinations, using simple chemical salts, and were compared with those of FYM. Eight of the original experiments still continue, some with little change and others appreciably modified. Because each of these experiments also has soils which have received no amendments, it is possible to assess the effects of man's anthropogenic activities on their fertility.

Data from these and other long-term experiments are used here to assess, where possible, effects of non-agricultural as well as agricultural inputs on the physico-chemical properties of soils and the above-ground biomass and its composition, and how these have changed with time.

6.1.1 SITES

Besides owning the Rothamsted Experimental Farm, the Lawes Agricultural Trust has a long-term tenancy on the Woburn Experimental Farm started in 1876 by the Royal Agricultural Society of England (Johnston, 1977). Rothamsted farm is about 330 ha (grid reference TL 130137). The farm lies in a 'semi-rural' location in Western Europe (Figure 6.1), about 42 km in a northerly direction from central London and within 2 km of two major trunk roads and a motorway (MI). The immediately surrounding area is primarily agricultural, although large urban and light industrial areas have been developed within 5-10 km since the 1950s. Woburn Experimental Farm is about 70 ha (grid reference SP 962360). The farm lies in a rural location about 27 km in a north-westerly direction from Rothamsted. The farm is adjacent to the Woburn Abbey Park, and the surrounding area is agricultural.

6.1.2 SOILS

The soils at Rothamsted and Woburn are silty clay loams and sandy loams respectively. In recent years average annual rainfall has been 700 and 640 mm. Both soils are free draining; those at Woburn have no free calcium carbonate, and to grow arable crops successfully regular dressings of chalk are needed (7.5 t/ha once every 6 years to maintain pH at about 6.5).

Figure 6.1 Location of Rothamsted, Woburn and Saxmundham Experimental Stations

The Rothamsted soil has developed in clay-with flints overlying chalk. The soil is naturally acid, but because the chalk is within 1 to 2 m of the surface in places, it was the practice in the eighteenth and early nineteenth centuries to dig and spread the chalk on the surface soil at rates up to 250 t/ha. However, this was expensive, and the chalk was applied only to fields growing arable crops, where it produced soils with pHs ranging from 7 to 8 and with reserves of free calcium carbonate. Grassland fields were not chalked and so remained acid. Soil pHs given here were measured in a 1 :2.5 soil: water suspension.

6.1.3 EXPERIMENTS

The experimental sites at Rothamsted are all on the same soil series and within 1-2 km of each other. The Park Grass experiment, which assessed the effects of manurial treatments on permanent grassland, started in 1856 on soil which was at pH 5.6 to 5.8. (Warren and Johnston, 1964). The Broadbalk experiment, in which winter wheat is grown each year, was started in 1843 on a soil with pH about 8 (Johnston and Garner, 1969). The Geescroft experiment started in 1847. Grain legumes were grown until 1878, the site was then fallowed for four years and clover grown from 1883 to 1885. Like Broadbalk, Geescroft was shown to be in arable cropping on an estate map of 1623 but it probably received less chalk than Broadbalk. Parts of both Broadbalk and Geescroft were fenced off in 1882 and 1886 respectively. Where the land has subsequently remained untended, it has reverted to deciduous woodland except where, on Broadbalk, one area is grazed each year by sheep, and another has the tree and shrub seedlings removed annually (Jenkinson, 1971 ). These experiments are now known respectively as the Broadbalk and Geescroft Wildernesses.

6.2 SOIL ACIDIFICATION

6.2.1 CHANGES IN SOIL pH DURING 100 YEARS

Johnston et al. (1986) recently summarized changes in soil pH for a period of more than 100 years in soil under permanent grassland on Park Grass (Figure 6.2(a)) and deciduous woodland in the Geescroft Wilderness (Figure 6.2(b)), and also assessed the relative contribution of various acidifying inputs. In 1856 the Park Grass soil had a pH of about 5.6 to 5.8. The unmanured soils acidified slightly between 1856 and 1876 and then showed a slight increase in pH from 5.4 (1876) to 5.7 (1923) because small dressings of chalk were applied between 1881 and 1896. From 1923, pH declined steadily and has now reached a predicted equilibrium value of pH 5.1 ± 0.2. The change in these soils, due to natural acidifying inputs, can be compared to that on plots receiving ammonium sulphate supplying 48, 96, or 144 kg N/ha each year. These soils have also reached a predicted equilibrium, at a pH of about 3.6. The soil given the largest amount of N had reached a pH of 3.6 by 1923 (Figure 6.2(a)); the other two soils have reached this value more slowly. Another soil receives sodium nitrate, supplying 96 kg N/ha each year, and its pH is now less acid than that of the unmanured soils (Figure 6.2(a)). One explanation could be that the added sodium, rather than calcium, leached accompanying anions in the drainage water. A similar but smaller effect of applying sodium nitrate was observed on the sandy loam soil at Woburn (Johnston and Chater, 1975). Differences in pH between otherwise similar soils in maritime and non-maritime climates could be explained by such an effect.

In 1886, the pH of the Geescroft Wilderness soil was probably just over 7. As deciduous woodland has developed, all three 23-cm horizons down to 69 cm have acidified (Figure 6.2(b)). While there was a delay in the onset of acidification in the lower horizons, acidification of subsoils was not delayed until the surface horizon reached its present low value. Both the 0 to 23 and 23 to 46 cm horizons have acidified more than the unmanured permanent grassland soil (compare Figures 6.2(a) and 6.2(b)) and are probably still acidifying. It appears that the woodland soils will have a pH nearer to that of grassland soils given ammonium sulphate (pH 3.6) than to that of the unmanured soil (pH 5.1). Thus, the pH of soil under regenerating deciduous woodland has fallen faster than that under grassland (2.9 units in 100 years compared with 0.7 units in 125 years) and to a lower equilibrium value (pH 3.6 compared with 5.1). The difference in equilibrium pH may be due to the trees being more efficient than grass at capturing wet and dry deposition. The acidifying inputs may now be much greater than previously, and could increase further.

Figure 6.2 (a) The pH values of soil samples taken from the unmanured (·), ammonium N (p) and nitrate-N (¾) treated plots of the Park Grass Experiment (Goulding et al., 1988); (b) the pH values of soil samples taken from the 0-23 cm (·),23-46 cm (¾) and 46-69 cm (p) horizons of Geescroft Wilderness (Goulding et al., 1988)

On the sandy loam soil at Woburn under continuous cereal cultivation, the pH of unmanured soils fell from 6.3 in 1876 to 5.3 in 1927 and then remained at this value for the next 27 years (Johnston and Chater, 1975). This equilibrium value of 5.3 is similar to that on the heavier textured soil under grassland at Rothamsted. This suggests that cereals and grassland differ little in their ability to trap aerial deposition, and more important, that the lime potential (pH -¹/₂p(Ca+Mg)) of the incoming rainfall had a greater effect on equilibrium soil pH than the texture of the soil (Johnston et al., 1986).

6.2.2 ACIDIFYING INPUTS

Johnston et al. (1986) also assessed the contribution to soil acidification of (1) soil-derived natural sources of acidity, (2) atmospheric deposition, (3) nutrient uptake by crops, and (4) fertilizer additions. Nutrient uptake is only important if crops are harvested and removed, although there is some retention in the standing crop under permanent vegetation. Fertilizer inputs need only be considered when applied in agricultural systems.

Soil-derived natural sources were assumed to include dissolution of soil-derived carbonic acid, nitrification of ammonia from mineralized organic matter, and the loss by leaching of base cations with sulphate and nitrate from organic matter. Atmospheric deposition included H⁺ deposition in rain, wet and dry deposition of oxides of sulphur and nitrogen, and nitrification of deposited ammonium. Wet deposition of H⁺, true acid rain, was calculated to be a negligible source of acidity at Rothamsted, contributing less than 10% of total atmospheric input and less than 2% of total inputs.

Goulding et al. (1988) have recently estimated acidifying inputs from both wet and dry sources as keq H⁺ /ha/yr at Rothamsted (Table 6.1 ). The potential acidity from the nitrification of NH⁺ in precipitation is becoming increasingly important. There are probably two main sources of this ammonia: volatilization from animal excreta, and the combustion of fossil fuels.

Table 6.1 Estimated amounts of wet and dry deposited H⁺ at Rothamsted (adapted from Goulding et al., 1988)

Some of the inputs of acidity have changed both in absolute amounts and relative importance during the course of both the Park Grass and Geescroft experiments (Table 6.2). The minimum H⁺ input was calculated from Bolton's (1977) equation relating calcium losses to pH. The amount of soil-derived natural sources is the difference between minimum H⁺ input and the sum of atmospheric deposition and nutrient uptake. Atmospheric deposition has increased, but the percentage contribution of soil-derived natural sources has decreased, probably because on acid soils the dissociation of carbonic acid is negligible, the nitrification of ammonia is very small, and there is less H⁺ - Ca²⁺ exchange (the calcium will be held more strongly and H⁺ will be buffered by aluminium). Thus, these two Rothamsted soils under woodland and grassland have settled at different equilibrium pH values, but those for the Rothamsted (grassland) and Woburn (arable) soils are similar. This suggests that the amount and intensity of aerial inputs are more important than soil buffering ability and texture, and that vegetative cover is important in trapping these inputs. The results also suggest that once the surface soil has reached its equilibrium value, all the atmospheric acidifying inputs are likely to be leached to acidify subsoils or to be transferred to aquatic ecosystems.

Table 6.2 Estimated amounts and relative contributions of the acidifying inputs to unmanured grassland and woodland (adapted from Goulding et al.. 1988). 

At these different equilibrium pH values, soil physico-chemical properties may well be different, soil microbial properties will be affected, and yield of above- ground biomass will differ.

6.3 EFFECT OF SOIL pH ON SOME SOIL PROPERTIES

Soil physico-chemical properties and microbial activity have considerable effects on nutrient losses from soil. In the l870s to l880s drainage from each of the Broadbalk winter wheat plots was collected and analyzed for the nutrients it contained. The data showed that where ammonium sulphate was applied, mainly nitrate was lost. The drainage contained large quantities of calcium but very little potassium even though this nutrient, rather than calcium, was applied as fertilizer. Very little phosphorus was lost.

6.3.1 FIXATION OF PHOSPHORUS

As soil acidity increases, phosphate is usually precipitated as iron and aluminium phosphates which are very insoluble (White, 1979). At low soil pH plants will become much more dependent on mycorrhizal associations and the cycling of phosphorus through soil organic matter. The benefit of inoculation with vesicular arbuscular mycorrhizal (VAM) fungi in the presence of rock phosphate on yields of cassava and sorghum grown on very acid, impoverished soils of Colombia has recently been demonstrated (Table 6.3). In the absence of rock phosphate, there was no yield of either cassava or sorghum. Similar benefits from combined inoculation with VAM fungi and Rhizobium have been reported for white clover grown on nutrient-poor, acidic soils in Wales (Table 6.4).

Table 6.3 Effect of rock phosphate and VAM inoculation (+M) on final yields of cassava and sorghum in Colombia (I. Arias, personal communication)

Table 6.4 Effect of inoculation with mycorrhizal fungi on dry matter production from upland swards at pH sod seeded with white clover³ (adapted from Hayman,1984)

6.3.2 RELEASE OF CATIONS

Potassium, calcium, and magnesium can all be released from soil minerals as soils acidify. This is shown indirectly in Table 6.5, where more of the potassium added to soil remained exchangeable in IN-ammonium acetate in acid soils than in neutral ones. Goulding and Stevens (1988) have estimated potassium reserves in acid upland soils under forest systems that were either clear-felled or whole-tree harvested. While the latter removed all nutrients, any returned under clear felling (trunks only removed) were poorly absorbed, and rapidly leached from this stagnopodzol at pH 4 with low cation exchange capacity and in a high-rainfall area. However, other data suggest that, at this particular forest at Beddgelert in North Wales, K-bearing minerals will weather quickly enough to supply the K requirements of conifers for tens of cycles of tree growth. The authors considered that other nutrients, like calcium and phosphorus, were more likely to become growth limiting.

Russell (1961) used unpublished results obtained by Schofield at Rothamsted to demonstrate that in acid soils, aluminium becomes an important exchangeable cation. Using a mildly acid (pH 5.6) and a very acid (pH 3.7) soil from Park Grass, Schofield showed that the titration curve relating soil pH and millequivalents of base taken up were identical for the very acid soil and the mildly acid soil after the latter had been pretreated with aluminium chloride. However the reverse process also held  -  the titration curves for the mildly acid and the very acid soils were similar after the latter had been pretreated first with dilute acid, to remove aluminium, and then with lime-saturated water to replace H⁺ by Ca²⁺.

Table 6.5 Effect of soil pH and past K manuring on the percentage of added K which remained exchangeable in soils which were alternately wet and dry for 12 weeks (adapted from Johnston, 1986)

Iron and aluminium, released in soluble form as soils acidify, may be transferred into aquatic ecosystems. Such transfers may well increase if land currently maintained at near neutral pH for arable food crops is set aside into forestry. The data in Figure 6.2(b) show that such land could become very acid over a period of years. The quantities of iron and aluminium released will depend on soil texture and composition of the clay.

6.3.3 EFFECTS OF LIMING

The extent and rapidity of change in pH of field soils depends on the degree of mixing of the neutralizing material. In Britain this is often chalk (calcium carbonate ) or lime (calcium hydroxide). In arable farming systems, mixing by ploughing and cultivation will often change pH throughout the plough layer within a year or so. In undisturbed soil, raising the pH to depth may take much longer. Table 6.6 shows amounts of chalk applied to Park Grass soils in attempts to change soil pH to either 5 or 6. In the presence of a mat of partially decomposed plant debris lying on the surface soil, the pH of the underlying mineral soil had not changed after six years. This was because the added calcium was held firmly on the cation exchange sites on the organic material. However, raising the pH of the mat increased biological activity and eventually the mat was decomposed. Once this had happened the calcium was leached into the mineral soil below and raised its pH.

Table 6.6 Effects of additions of calcium carbonate on soil pH (adapted from Johnston, 1972)

6.4 EFFECT OF SOIL pH ON FLORA

6.4.1 EFFECT ON SOIL FLORA

Examples of complex interactions between soil nutrient status, pH, and beneficial soil organisms, like VAM fungi and Rhizobium, have been discussed earlier in this chapter. However, soil acidity can also affect pathogenic organisms. Mary D.Glynne surveyed fields on Rothamsted and Woburn farms for the incidence of various cereal diseases including take-all, caused by Gaeumannomyces graminis, a fungus attacking the stem base. She found very little take-all on soils at pH 5, but considerably more at higher pHs (Glynne, 1935).

6.4.2 EFFECT OF SOIL REACTION ON ABOVE-GROUND  BIOMASS AND ITS COMPOSITION

It is well established that plants differ in their tolerance to varying soil reaction. Differences between the Broadbalk and Geescroft Wildernesses and within the Park Grass experiment are good examples because climate and soil type are identical and the effects observed must be due to differences in pH or nutrient status.

The causes and extent of acidification at Geescroft have been discussed above, while mention was made that the Broadbalk soil has been buffered at about pH 8 for nearly 200 years. The effects of the large difference in pH are clearly seen in variations in stand composition and dry weight, which are smaller on Geescroft than on Broadbalk (Table 6.7). Geescroft is almost entirely composed of oak (58% ) and ash (26%), with some hawthorn (8%) and elm (5%). Broadbalk has ash (38%), hawthorn (32%), sycamore (18%), and some maple (8%). Both ash and maple are less tolerant of acid soils and the long-term decline of pH on Geescroft could affect their ability to compete and regenerate, making oak even more dominant. However, some recent unpublished evidence suggests that there may be inhibition of regeneration of oak at low pH.

The yield and species composition of the Park Grass plots strikingly demonstrate both the effects of acidity and of ameliorating that acidity with dressings of chalk since 1903. (For details of chalking see Warren and Johnston, 1964, and for species changes Thurston et al., 1976.) Absolute yield levels, but not the effect of treatment, have been affected by a change in the method of harvesting from 1960 (Table 6.8). Before 1960, hay was made on the field in late June with attendant dry matter losses during haymaking. Since then herbage has been cut green, weighed, and dry matter determined immediately; yields were therefore apparently increased. On the unmanured plots, the small differences in yield between limed (pH 7.1) and unlimed (pH 5.1) sections suggests that lack of nutrients, mainly nitrogen, rather than acidification was the more important determinant of yield. On plots receiving ammonium sulphate (144 kg N/ha), there was a large benefit from lime; in the absence of lime, pH fell to 3.6, and yield was considerably diminished.

Table 6.7  Stand measurements (dry weight in t/ha) of Broadbalk and Geescroft Wilderness sites (adapted from Jenkinson, 1971)

Table 6.8 Effect of acidity on the yield of herbage from the Park Grass Experiment (adapted from Goulding et al., 1988)

Table 6.9  Effect of acidity on the botanical composition of the permanent pasture of Park Grass (adapted from Goulding et al., 1988) 

Table 6.10 Effect of manuring and soil reaction on grass species, Park Grass, by 1947-49 (adapted from Warren and Johnson, 1964)

Each species expressed as per cent by weight of the grass fraction; species present in amounts less than 10% are omitted from the table

Soils receiving sodium nitrate had a higher pH than unmanured soils (Figure 6.2(a)), and yields were as good as those on soils receiving ammonium sulphate plus lime. The same comparisons for botanical composition (Table 6.9) suggest that there is a complex interaction between pH and nutrition (for further details see Thurston et al., 1976). Supplying N greatly increased yield (Table 6.8) but also enormously decreased species diversity (Table 6.9). Species composition can be expressed as the proportion of each species as a per cent by weight of each group: grasses, legumes, or herbs. For the various grass species this is shown in Table 6.10. Not all combinations of pH and nutrient effects are available because not all exist in the current plot treatments. On soils with pH values 3.7 to 4.1 and where the manuring was deficient in phosphorus, the main grass was Agrostis tenuis. When P was given together with ammonium nitrogen, Agrostis tenuis was still the dominant species but the grass fraction of the herbage contained about 20% of Holcus lanatus. With complete NPK manuring, all or nearly all of the grass was Holcus lanatus. With the same complete manuring, but on less acid soils at pH 4.2 to 6.0, this grass was replaced chiefly by Alopecurus pratensis and small amounts of Arrhenatherum elatius. Omitting K decreased the Alopecurus pratensis on these soils (pH 4.2 to 6.0) and Festuca rubra became the dominant grass, except where only 48 kg N/ha (as sodium nitrate, Plot 17) was given, and Dactylis glomerata then replaced Alopecurus pratensis. In the third group of soils (pH 6.0 to 7.5), where P and K were omitted, Alopecurus pratensis was replaced by Dactylis glomerata, Festuca rubra, and Heliotrotrichon pubescens. With complete NPK (Plot 14 ), Alopecurus pratensis was present but there was also much Arrhenatherum elatius and Dactylis glomerata. The results showed that Alopecurus pratensis needs P and K in addition to N fertilizer, and that there is a critical limit of soil acidity (pH 4.2) below which it does not grow; but from these data less is known about its growth in neutral and slightly calcareous soils.

While the above results on the effects of changing pH and nutrient status on yield and competition between species are themselves interesting, recent work suggests that we may not always be dealing with the same genetic population. Anthoxanthum odoratum has always been one of the most widely distributed but not necessarily dominant species, occurring in both limed and unlimed, fertilized and unfertilized plots. Recently, it has been one of the most abundant species on some of the more acid soils. Its wide distribution could be due to the fact that morphologically and physiologically different populations have evolved on the various plots. Recent information, discussed by Thurston et al. (1976), suggests that there have been significant changes within the species. Morphological differences exist between the populations and are apparently adaptive. The differences are usually correlated with environmental conditions on the plots from which they were collected (Snaydon, 1970; Snaydon and Davies, 1972).

Populations also differ in their response to mineral nutrients; again the differences appear to be adaptive (Davies and Snaydon, 1973a,b, 1974; Davies, 1975). The adaptive nature of the differences between populations was confirmed by growing them under uniform conditions for several years and then transplanting them back into either their own or other plots. Each population survived and grew fastest on its own plot (Davies and Snaydon, 1976).

Similar studies of Anthoxanthum odoratum populations from subplots which were not limed until 1965 indicate that genetic changes have occurred within seven years, and that populations collected less than 1 m apart from opposite sides of plot boundaries differ both morphologically and physiologically (Snaydon and Davies, 1976). These differences have developed despite appreciable gene flow caused by pollen drift and seed dispersal. Similar differences have also been demonstrated between populations of Lolium perenne (Goodman, 1969), Holcus lanatus, and Dactylis glomerata (Remison, 1976).

6.5 OTHER INPUTS TO SOIL

Many industrial processes release compounds into the atmosphere in particulate or gaseous forms, and these may eventually be deposited on soil. Improvements in analytical techniques now allow us to investigate such occurrences by analyzing crops and soils from the Rothamsted archive which dates back to 1843, an excellent reason for maintaining such an archive.

6.5.1 TEMPORAL CHANGES IN CADMIUM CONTENT

Recently, soils from Park Grass and three of the Classical experiments growing arable crops have been analyzed to investigate time trends in cadmium content due either solely to atmospheric deposition or a combination of atmospheric deposition and various soil treatments (Jones et al., 1987). Since the 1850s, there has been an increase in cadmium concentration in the soil plough layer of between 0.7 and 1.9 µg/kg, equivalent to an increase of 1.9 to 5.4 g Cd/ha/yr, due to atmospheric deposition. The changes in soil cadmium concentrations since 1846 at one site corresponded well to predicted increases in the plough layer cadmium burden based on assumptions about the temporal trends in atmospheric cadmium emissions (Figure 6.3).

In addition, subsamples of a selection of rock phosphates of known origin and superphosphates mainly from one supplier, collected and stored in the archive since 1925, were also analyzed for cadmium. The concentrations ranged from 3.6 to 92 (mean 36) mg Cd/kg for rock phosphates and from 3.3 to 40 (9.7) mg Cd/kg for superphosphates. On the basis of these data and their known application rates, the estimated annual input of cadmium to P-treated plots at Rothamsted was 2 g/ha. However, there was little further increase in soil cadmium due to this addition in the three arable experiments where soil pH was >6.5. On these P-treated plots, the mean increase in soil cadmium was 1.2 g/kg/yr which is equivalent to an increase in the plough layer burden of 3.1g Cd/ha/yr. This suggests there may have been some loss by leaching and highlights the need for comprehensive balance studies. The results also raise another question, namely has cadmium in aerial inputs behaved differently from that in superphosphate? Is there a difference in speciation or are the differences due to only limited adsorption sites at the level of organic matter in these soils? By contrast, P-treated soils under permanent grassland, with a higher organic matter content and a lower pH, have increased their cadmium content by 7.2 g/ha/yr. When permanent grassland soils ranging in pH from 5 to 7 were examined, it was found that organic matter had a larger effect on cadmium concentration than pH, and the effects of pH were not consistent (Jones et al., 1987).

Figure 6.3 Change in soil cadmium concentration with time on the unmanured soil on Broadbalk. Measured data (X-X), predicted soil Cd concentrations (O-O). Error bars relate to analyses of eight separately digested samples for both 1846 and 1980 (Jones et al., 1987)

Data in Figure 6.3 show that there has been a very large increase in soil cadmium burden since the 1940s, and other results indicate that there has been some retention of this cadmium in soils with large amounts of organic matter. Jones et al. (1987) also observed that farmyard manure, applied to some experimental plots at Rothamsted, appeared to have been a more significant source of cadmium than combined atmospheric and phosphate fertilizer inputs. Larger inputs may have been responsible; for example, McGrath (1984) reported values of 1.8 g Cd/t dry FYM, but it may be that the increased organic matter content of the soil, as a result of adding farmyard manure (35 t/ha/yr) over a period of more than one hundred years, has strongly held the cadmium applied in FYM or has retained more of the aerial inputs against leaching.

Cereal grain and herbage samples in the archive have been analyzed for their cadmium content because, in part, the significance of the long-term increases in soil cadmium depend on whether increases in plant cadmium concentration are detectable (Jones and Johnston, 1989b). There was little evidence (with one exception) of a long-term increase in grain cadmium concentrations at Rothamsted. Herbage removed from Park Grass has always had a larger concentration of cadmium than cereal grain, approximately 200 and 40 µg/kg dry weight respectively. The herbage was not washed prior to analysis and its larger cadmium concentration may be due to surface contamination. However, the values are probably too large for this to be the sole reason. Wherever the cadmium is located, it will be a source of dietary cadmium to animals.

6.5.2 TEMPORAL CHANGES IN POLYNUCLEAR AROMATIC HYDROCARBONS

The changing levels of polynuclear aromatic hydrocarbons (PAHs), which are mutagenic and carcinogenic, have been investigated. The total PAH burden of the plough layer has increased approximately fivefold since the 1880s to 1890s with some compounds showing substantially greater increases (Jones et al., 1989d). Average rates of increase for individual PAHs over the century since 1880 to 1890 vary between 0.01 to 0.67 (mean 0.21) mg/m²/yr. These fluxes are similar to contemporary atmospheric deposition rates at other semi-rural locations in Britain. This fall-out of anthropogenically generated PAHs is derived from combustion of organic materials. What is of concern is the observed large increase in soil PAH content in the latter half of this century which may well be representative of other soils in the industrialized countries or regions (Figure 6.4). The similarities between the average annual rate of increase in the soil PAH burden and the likely average deposition flux at Rothamsted suggest that losses via five possible mechanisms (microbial breakdown, photo-oxidation, vaporization, crop offtake, and leaching) effectively remove only relatively small proportions of the total annual input. This implies long residence times for PAHs in soils.

In general, compounds with complex structures have increased more than those with simpler ones, suggesting that microbial breakdown and soil retention of PAHs may well depend on molecular structure. This poses the question of whether some of these more complex compounds may be toxic to the soil microbial population. Clearly there is a need to quantify both inputs to, and outputs from, sites of ecological importance if it can be shown that these materials have adverse effects on soil microbial processes.

There has been no change in the PAH concentration of wheat and barley grains at Rothamsted over many years. This suggests that either plants can exclude PAH uptake at the root surface or they do not translocate them into seeds. Varying amounts of PAHs in herbage samples from Park Grass suggest that this may be surface contamination (Jones et al., 1989a). For herbivores, dietary intake may arise from these surface deposits. Soil profile samples collected from the same plots on Broadbalk in 1893, 1944, and 1987 have been analyzed for their PAH content (Jones et al., 1989c). The total PAH content in the 1893 samples showed little enrichment of surface soil relative to that in the subsoil. The values were similar to those in contemporary isolated rural locations in Britain. By 1987, the surface soil had been enriched in all PAH compounds measured by a factor of between 1.3 (acenaphthalene) and over 20 (benzo[a]pyrene). Increases in the PAH content of the 23- to 46-cm sub-surface layer indicated some migration of PAHs from the plough layer. Net average annual migration rates range from 0.01 to 0.14 mg/m² for individual PAHs, and the rate appeared to be primarily a function of the plough layer PAH content rather than the physical/chemical properties of the individual compounds. This suggests that particle-bound translocation is the dominant mechanism for PAH migration. Such movement through soils or by erosion into rivers has implications deserving further study if these compounds affect plants or animals in aquatic ecosystems.

Figure 6.4 Change in soil PAH concentration with time on the unmanured soil on Broadbalk (Jones et al., 1989d)

Figure 6.5 Modelled total (wet plus dry) sulphur deposition (a) for l970 and (b) for 1983 (gSm^-2 year ^-1) (Warren Spring Laboratory, 1985)

Figure 6.6 (a) Wet and dry sulphur deposition and (b) sulphur in wheat grain in England,  Wales and Scotland (McGrath and Johnson, 1986)

6.5.3 SULPHUR CYCLING

The changing patterns of man's anthropogenic activities is seen also in the cycling of sulphur. There has been outcry that sulphur emissions, from the combustion of fossil fuels, are acidifying inputs which can rapidly acidify some soils. As a result of changing technology, atmospheric deposition of sulphur to many soils in the UK has declined rapidly (Figure 6.5). At the same time, other sulphur inputs to agricultural soils have also diminished (McGrath and Johnston, 1986). Much sulphur used to be added in either single superphosphate or ammonium sulphate, two fertilizers which are now no longer generally available in the UK, and little sulphur is added in other fertilizers. Because both of these sulphur inputs have declined there is now concern about the possible need to apply sulphur to soils growing arable crops with a high sulphur demand, like oilseed rape. Perhaps more important may be the need to supply sulphur to maintain protein quality of cereal grains (Figure 6.6) for bread making and for those people dependent on grain legumes for their protein intake.

6.6 THE EFFECT ON SOILS OF CHANGING LAND USE

One response to the surplus of agricultural products within Western Europe has been the suggestion that land should be taken out of farming to revert to other use. Reference has already been made to how deciduous woodlands have regenerated at Rothamsted when arable soils are totally unmanaged and the effect on soil pH, and, in its turn, that of pH on biomass production and species composition. In addition, the soils have accumulated large, but different, amounts of organic matter, sulphur, and phosphorus. Build-up in the Geescroft Wilderness soil, where the pH has become gradually more acid, is much less than in the Broadbalk soil (Table 6.11 ). This may be because inputs to the Geescroft site have been smaller than those on Broadbalk and/or decomposition has been slower on Broadbalk. However, the difference between these two sites, which are in close proximity, is so large that there is a need for a greater understanding of the processes which have enhanced accumulation of organic matter, and therefore fertility, in one soil rather than the other. The one factor which is explicable is the greater gain of inorganic sulphur in the Geescroft soil, which is due to the larger number of positive, pH-dependent absorption sites on a soil with low pH.

Table 6.11 Mean annual gains (kg/ha/yr) in organic C, N, S and P and in inorganic S in the topsoil (0-23 cm) of Broadbalk and Geescroft Wilderness soils (adapted from Jenkinson, 1971)   

Figure 6.7 Change in the total nitrogen content of the top 23 cm of soil under wheat either unmanured or receiving 35 t/ha each year and in the section of the Broadbalk Wilderness from which tree seedlings are removed each year (Jenkinson, 1971)

An interesting comparison can be made between the change in total nitrogen content in the top 23 cm of soil under the Broadbalk Wilderness (pH about 8), where tree seedlings have been removed each year, and that on the rest of the field where wheat is grown year after year (Jenkinson, 1971). Figure 6.7 shows that under continuous wheat the total nitrogen content of the unmanured soil has remained approximately constant over the last hundred years. Where FYM (35 t/ha) has been applied annually the nitrogen content has increased appreciably and is now near equilibrium. However, at the same time much nitrogen has been lost, some by crop uptake and some by leaching. Johnston et al. (1989) have recently calculated that about 125 kg N/ha/yr has been lost during the course of the experiment, probably as nitrate by leaching. The amount of total nitrogen which has accumulated in the Broadbalk Wilderness soil during the 100 years since 1886 is about as much as has accumulated in the soil given FYM each year. Whether the soil under woodland will have the same equilibrium organic matter content as that growing wheat and given FYM is not known, but if it is the same then there is a risk that nitrate could leach from soils once equilibrium has been reached.

6.7 CONCLUSIONS

Current funding strategies in biological research tend to favor short-term projects. However, long-term experiments are needed to quantify the effects of the many changes which affect soil fertility over the 25- to 100-year time scale. Funding agencies often disregard or do not appreciate these needs.

A good example of the short- and long-term approach is work on soil microbial biomass and organic matter at Rothamsted. Current results show that the microbial biomass fraction of soil organic matter changes proportionally more rapidly than does total soil organic matter (Powlson and Brookes, 1987). Changes in amounts of soil microbial biomass can be used to predict the direction of change in organic matter as a result of changes in management. However, they cannot be used to predict the equilibrium value at which total soil organic matter will settle. It is not until appreciable differences in total soil organic matter have been established that its effect on soil productivity can be measured (see Johnston, 1990). Another example, one of the inability of short-term observations to predict long-term changes in soil pests, was given by Johnston (1989).

This paper summarizes some further examples illustrating the importance of long-term experiments and others are given by Johnston (1990). All the examples have been taken from data accumulated at Rothamsted since 1843, Woburn since 1876, and Saxmundham since 1899. These experiments were mainly concerned with managed agricultural systems. However, some treatments allow the measurement of effects under less managed conditions. For example, we can estimate the effects of man's anthropogenic activities on soils which receive no agricultural treatment.

The results discussed here can be strictly applied only to similar soils under similar climatic conditions and farming systems. One of the most difficult problems is to assess the general applicability of the results. The danger of assuming similar effects on all soils is well illustrated above, where effects of acidifying inputs on tree growth are discussed. Johnston (1990) also showed very different effects of ammonium sulphate on spring barley yields at the Rothamsted and Woburn sites. The effects were due to differences in soil texture and initial soil pH. Results from similar long-term research would be invaluable, and, indeed, are essential to the production and validation of models to describe on a global scale some of the underlying chemical and biological processes occurring in soil.

The Rothamsted experiments have a number of important features worth careful consideration by research workers planning, and managers funding, such long-term research. The experiments are the responsibility of the Lawes Agricultural Trust Committee, which can continue in perpetuity. There is thus security of tenure of both the Rothamsted and Woburn sites which has been vital not only to the continuation of the long-term experiments but also to the realization by staff that their commitment to a long-term research program would not suddenly stop because the site was no longer available.

The scientific management of the experiments is delegated to a group of scientists rather than an individual. Each group is multidisciplinary and must refer major management changes to the Trust Committee for their approval. The multidisciplinary approach ensures that data are collected for a wide range of factors and that many possible interactions are studied.

A unique feature of the Rothamsted experiments is the archive of crop and soil samples. Samples of harvested cereal crops, grain and straw, and herbage exist for most years and treatments from the start of the experiments. Soil samples have also been taken periodically. Lawes and Gilbert sampled by 9-inch depth horizons, although in the 1840s to 1860s cultivation would not have exceeded 4 to 5 inches. This has proved to be a major benefit because plough depth has since been increased to about 9 inches. The constancy of sampling depth means that comparisons are possible for total soil constituents even though plough depth has changed.

The foresight by Lawes and Gilbert highlights a major need in new experiments where changes in soil properties must be an important factor. At the outset, a soil sampling protocol must be established and then adhered to. The interval between taking samples is less important, and can be related to the length of a crop rotation in agricultural experiments, but it should probably not be longer than every 10 years. Both crop and soil samples should be dried and stored under conditions which prevent deterioration.

Current analytical techniques now allow the opportunity to measure the concentration of a wide range of inorganic and organic constituents. One reason for maintaining archived samples is to allow other constituents to be determined or past analyses updated using more accurate methods. We cannot improve on Lawes and Gilbert's data for total nitrogen in soil, but we can now analyze for zinc, copper, nickel, and cadmium, and polynuclear aromatic hydrocarbons whose significance was not appreciated in the last century .

Such long data sets are essential to estimate small but consistent long-term changes in soil composition, and their effects on crop growth and composition. Without such data sets it is impossible to discuss with conviction the effects of agriculture on the environment, and man's anthropogenic activities on the soil. In the final analysis, soil is one of a country's most important natural assets, on which the ability to feed its population, or to provide them with an acceptable landscape, depends. This chapter has attempted to show by examples that long-term fertility trends in soil can be measured and understood only by using data from long-term experiments.

6.8 REFERENCES

Bolton J. (1977). Changes in soil pH and exchangeable calcium in two liming experiments on contrasting soils over 12 years. Journal of Agricultural Science, Cambridge, 89, 81-86.

Davies, M.S. (1975). Physiological differences among populations of Anthoxanthum odoratum L. collected from the Park Grass Experiment, Rothamsted. IV. Response to potassium and magnesium. J. of Appl. Ecol., 12, 953-964.

Davies, M.S. and Snaydon, R.W. (1973a) Physiological differences among populations of Anthoxanthum odoratum L. collected from the Park Grass Experiment, Rothamsted. I. Response to calcium. J. of Appl. Ecol., 10, 33-45.

Davies, M.S. and Snaydon, R.W. (1973b). Physiological differences among populations of Anthoxanthum odoratum L. collected from the Park Grass Experiment, Rothamsted. II. Response to aluminium. J. of Appl. Ecol., 10, 47-55.

Davies, M.S. and Snaydon, R.W. (1974). Physiological differences among populations of Anthoxanthum odoratum L. collected from .the Park Grass Experiment, Rothamsted. III. Response to phosphate. J. of Appl. Ecol., 11, 699-707.

Davies, M.S. and Snaydon, R.W. (1976). Rapid population differentiation in a mosaic environment. III. Measurements of selection pressures. Heredity, 36, 59-66.

Glynne, M.D. (1935). Incidence of take-all on wheat and barley on experimental plots at Woburn. Ann. Appl. Bioi., 22, 225-235.

Goodman, P.J. (1969). Intra-specific variation in mineral nutrition of plants from different habitats. In Rorison, I.H. (Ed.) Ecological Aspects of the Mineral Nutrition of Plants . Blackwell Scientific, London, 237-253.

Goulding, K.W.T., Johnston, A.E. and Poulton, P.R. (1988). The effect of atmospheric deposition, especially of nitrogen, on grassland and woodland at Rothamsted Experimental Station, England, measured over more than 100 years. In Mathy, P. (Ed.) Air Pollution   and Ecosystems. Proceedings of an EEC International Symposium, Grenoble, France, May 1987. D. Reidel, Dordrecht, 841-846.

Goulding, K.W.T. and Stevens, P.A. (1988). Potassium reserves in a forested, acid upland soil and the effect on them of clear-felling versus whole-tree harvesting. Soil Use and Management, 4, 45-51.

Hayman, D.S. (1984). Improved establishment of white clover in hill grasslands by inoculation with mycorrhizal fungi. In Thomson, D.J. (Ed.) Forage Legumes. British Grasslands Society Occasional Symposium No.16. BGS Press, Hurley, Maidenhead, 44- 47.

Jenkinson, D.S. (1971). The accumulation of organic matter in soil left uncultivated. Rothamsted Experimental Station Report for 1970, Part 2, 113-137.

Johnston, A.E. (1972). Changes in soil properties caused by the new liming scheme on Park Grass. Rothamsted Experimental Station Report for 1971, Part 2, 177-180.

Johnston, A.E. (1977). Woburn Experimental Farm: a hundred years of agricultural research devoted to improving the productivity of light land. Lawes Agricultural Trust, Harpenden.

Johnston, A.E. (1986). Potassium fertilization to maintain a K-balance under various farming systems. In Nutrient Balances and the Need for Potassium. 13th Congress of the International Potash Institute, Reims, France, 177-204.

Johnston, A.E. (1989). The value of long-term experiments - a personal view. In Likens, G.E. (Ed.) Long-Term Studies in Ecology: Approaches and Alternatives. Springer-Verlag, New York, 175-179.

Johnston, A.E. (1990). The value of long-term experiments in agricultural research. Proceedings of the centennial of the Sanborn Plots, July 1988, University of Missouri-Columbia.

Johnston, A.E. and Chater, M. (1975). Experiments made on Stackyard Field, Woburn, 1876-1974. II. Effects of the treatments on soil pH, P and K in the Continuous Wheat and Barley experiments. Rothamsted Experimental Station Report for 1974, Part 2, 45-60.

Johnston, A.E. and Garner, H.V. (1969). The Broadbalk Wheat Experiment: Historical Introduction. Rothamsted Experimental Station Report for 1968, Part 2, 12-25.

Johnston, A.E. and Goulding, K.W.T. (1988). Rational K manuring. Journal of the Science of Food and Agriculture, 43, 319-320.

Johnston, A.E., Goulding, K.W.T. and Poulton, P.R. (1986). Soil acidification during more than 100 years under permanent grassland and woodland at Rothamsted. Soil U se and Management, 2, 3-10.

Johnston, A.E., McGrath, S.P., Poulton, P.R. and Lane, P.W. (1989). Accumulation and loss of nitrogen from manure, sludge and compost: long-term experiments at Rothamsted and Woburn. In Hansen, J.A. and Henriksen, K. (Eds) Nitrogen in Organic Wastes Applied to Soils. Academic Press, London, 126-139.

Jones, K.C., Grimmer, G., Jacob, J. and Johnston, A.E. (1989a). Changes in the polynuclear aromatic hydrocarbon (PAH) content of wheat gram and pasture grassland over the last century from one site in the U.K. Science of the Total Environment, 78, 117-130.

Jones, K.C. and Johnston, A.E. (1989b). Cadmium in cereal grain and herbage from long-term experimental plots at Rothamsted, U.K. Environmental Pollution, 57, 199-216.

Jones, K.C., Stratford, J.A., Tidridge, P., Waterhouse, K.S. and Johnston, A.E. (1989c). Polynuclear aromatic hydrocarbons in an agricultural soil: long-term changes in profile distribution. Environmental Pollution, 56, 337-351.

Jones, K.C., Stratford, J.A., Waterhouse, K.S., Furlong, E.T., Giger, W., Hites, R.A., Schaffner, C. and Johnston, A.E. (1989d). Increases in the polynuclear aromatic hydrocarbon (PAH) content of an agricultural soil over the last century. Environmental Science and Technology, 23, 95-101.

Jones, K.C., Symon, C.J. and Johnston, A.E. (1987). Retrospective analysis of an archived soil collection. II. Cadmium. The Science of the Total Environment, 67, 75-89.

McGrath, S.P. (1984). Metal concentrations in sludges and soil from a long-term field trial. Journal of Agricultural Science, 103, 25-35.

McGrath, S.P. and Johnston, A.E. (1986). Sulphur-crop nutrient and fungicide. Span (Progress in Agriculture), 29, 57-59.

Powlson, D.S. and Brookes, P.C. (1987) Measurement of soil microbial biomass provides an early indication of changes in total soil organic matter due to straw incorporation. Soil Biology and Biochemistry, 19, 159-164.

Remison, S.U. (1976). A study of root interactions among grass species. PhD Thesis, University of Reading.

Russell, E.W. (1961). Soil Conditions and Plant Growth. 9th edition. Longman, London. 

Snaydon, R.W. (1970). Rapid population differentiation in a mosaic environment. I. Response of Anthoxanthum odoratum L. populations to soils. Evolution, 24, 257-269. 

Snaydon, R.W. and Davies, M.S. (1972). Rapid population differentiation in a mosaic environment. II. Morphological variation in Anthoxanthum odoratum L. Evolution, 26, 390--405.

Snaydon, R.W. and Davies, M.S. (1976). Rapid population differentiation in a mosaic environment. IV. Populations of Anthoxanthum odoratum L. at sharp boundaries. Heredity, 37, 9-25.

Thurston, J.M., Williams, E.D. and Johnston, A.E. (1976). Modem developments in an experiment on permanent grassland started in 1856; effects of fertilizers and lime on botanical composition and crop and soil analyses. Annales Agronomiques, 27 (5-6), 1043- 1082.

Warren, R.G. and Johnston, A.E. (1964). The Park Grass Experiment. Rothamsted Experimental Station Report for 1963, 240-262.

Warren Spring Laboratory (1985). Acid Deposition in the United Kingdom. Second Report of the United Kingdom Review Group on Acid Rain. Department of Trade and Industry, London.

White. R.E. (1979). Introduction to the Principles and Practice of Soil Science. Blackwell, London.