Cumulative compaction of a clay loam soil by annually repeated field traffic in autumn

The cumulative effects of annually repeated field traffic on soil properties and barley yield were investigated in a field experiment on clay loam. Experimental traffic was applied with a tractortrailer combination prior to autumn ploughing for four successive years. The trailer single axle load was 5 Mg, The loading intensity was 0, 100 and 300 Mg km ha ', and both standard and low-profile trailer tyres were used. The effect of early summer irrigation on the yield was also studied. The yield and nitrogen uptake of the crop were determined for four successive years. Soil penetrometer resistance was measured annually after the second loading. The traffic compacted the soil to 0.35 m depth. On average, soil compaction reduced barley yield by 5% and nitrogen uptake by 7%. No annual cumulative increase in the compaction depth or yield reductions was found. Probably only the first loading compacted the subsoil, because the soil was drier than field capacity in the 0.2-0.3 m layer in the following autumns. The use of trailer lowprofile tyres did not reduce the depth of compaction or yield losses. On average, early summer irrigation increased grain yield by 34% and nitrogen uptake by 25%, but it did not significantly decrease yield or nitrogen uptake reductions due to compaction.


ntroduction
Harvesting time is short in Finland and field work often has to be done on moist soil.The tyre inflation pressure of combines and trailers is usually more than 100kPa and increasingly their axle load exceeds 5 Mg.Thus, there is a risk of subsoil compaction.Field traffic with a single axle load of 6.4-12 Mg and a tyre inflation pressure of 50-300 kPa is found to compact moist mineral soils to depths of 0.3-0.8m (Voorhees et al. 1986, van den Akker 1988, Lowery and Schuler 1991, Danfors 1994).Under unfavourable soil conditions, even an axle load of less than 5 Mg compacts the subsoil.Aura (1983) found that spring traffic with an axle load of 3 Mg and tyre inflation pressure of 140kPa compacted wet clay soil under the 0.2 m plough layer.
Severely compacted soil may take several years to recover.Despite annual ploughing and frost, severe compaction persists in the plough layer of clay soils for three (Alakukku 1996 a) or even five years (Håkansson and Danfors 1981).Regular tillage does not loosen the sub- soil, and the alleviation of subsoil compaction is usually left to natural processes.Subsoil com- paction may thus often persist for a long time.After a single heavy loading in soils with clay contents of 6-85% it has been reported to persist for 3-11 years despite cropping and deep frost (Blake et al. 1976, Voorhees et al. 1986, Håkansson 1985, Alakukku 1996 b).
Since alleviation of severe compaction takes so long, heavy loading repeated in the same place each year may increase soil compactness and yield losses year by year.Arvidsson and Håkan- son (1994) reported that annually repeated load- ing by 350 Mg km ha' 1 on wet clay soils before ploughing caused cumulative yield losses of spring cereals during the first four years.There- after the yields reached equilibrium.Gameda et al. (1987Gameda et al. ( ,1994) ) did not observe cumulative yield losses with maize.Field traffic with axle loads of 10-20 Mg and tyre inflation pressures > 300 kPa increased, however, the maximum soil density and the depth to which compaction occurred each year during the first three years (Gameda et al. 1987).Repeated field traffic does not, how- ever, always increase soil compaction cumula- tively.Alblas et al. (1994) loaded sandy soils with an axle load of 10Mg twice a year during a period offour years, and did not find any cumu- lative soil compaction or cumulative silage maize yield losses.
The introduction of large-volume, low-profile tractor and implement tyres allows inflation pressures of 50-100 kPa, even with heavy loads.The use oflow ground contact pressure machines is expected to allow field traffic with quite heavy axle loads.Danfors (1994) reported, however, that a reduction in inflation pressure from 150 to 50 kPa reduced the compaction of moist clay soils with axle loads > 8 Mg to a maximum depth of 0.3-0.4m only.This finding is consistent with theoretical models and earlier measurements showing that the average ground contact pressure has the greatest influence on soil compact- ion in the topsoil and upper part of the subsoil, but that the axle load is more important deeper in the subsoil (Söhne 1958, Danfors 1974, Bol- ling 1987, Olsen 1994).Arvidsson and Håkansson (1994) found that the use of low-profile trailer tyres in their trac- tor-trailer combination (350 Mg km ha' 1 ), even at an inflation pressure as high as 200 kPa, reduced the yield of spring cereals by 5 percent- age units less than standard tyres (500 kPa) as a mean of four years.Vermeulen and Perdok (1994) compared a full-size low ground contact pressure farming system with tyre inflation pressures of 40-80 kPa and a common ground pressure system with inflation pressures of 80-240 kPa.On clay loam, the ground pressure did not influence the wheat yield as a mean offour years (Vermeulen and Perdok 1994).Likewise, Chamen et al. (1990) found that winter wheat yields were similar in plots with low (< 50 kPa) and high (100-250 kPa) ground contact pressure.
Soil mechanical resistance may prevent roots from penetrating compacted soil when the soil dries up.Some studies indicate that irrigation may relieve this harmful effect.In a year with low precipitation early in the growing season, Elonen (1980) found that irrigation four weeks after sowing reduced yield losses of spring wheat due to clay soil compaction by a 3 Mg axle load just before sowing.Irrigation softened the dry soil and improved root penetration.Likewise, irrigation reduced yield loss of winter wheat in plots with a plough pan (Barraclough and Weir   1988).
In earlier Finnish field experiments, traffic on a single occasion with a tandem axle load of 16 Mg and a tyre inflation pressure of 700 kPa compacted moist clay soil to 0.5 m depth, and subsoil compaction persisted for at least nine years (Alakukku 1996 b).During the harvest, however, field traffic with 5-10 Mg axle loads is repeated each year.Likewise, low-profile tyres allow the tyre inflation pressure to be reduced without decreasing the loading capacity of the tyre.In this study, a field experiment was con- ducted on clay loam objectives to (1) investigate the cumulative effects of annual transport traffic prior to autumn ploughing on soil and on 446 AGRICULTURAL SCIENCE IN FINLAND barley yield, (2) evaluate the effect of the infla- tion pressure of trailer tyres on soil compaction, (3) determine the subsoil compaction by a load of no more than 5 Mg on a single axle, and (4) establish whether there is an interaction between compaction and irrigation in early summer, when precipitation is often low in southern Finland.

Experimental field and treatments
The field trial was performed at the Agricultural Research Centre of Finland at Jokioinen (60"49'N, 23°28'E) on a clay loam soil (Table 1, Vertic Cambisol (FAO 1988)) in 1985-1989.The experiment was laid out in a randomized com- plete-block design with a strip-plot arrangement and four replicates.The main plot treatment (A) was sprinkler irrigation in June with two factor levels: (1) control, no irrigation, and (2) irrigation with 20-30 mm once or twice.The subplot treatment (BC) was field traffic with a tractortrailer combination in autumn with five factor levels: (1) control, no experimental traffic, (2) 100 Mg km ha 1, low-profile (LP) tyres on the trailer (inflation pressure 150 kPa), (3) 100 Mg km ha' 1 , standard (ST) tyres on the trailer (350 kPa), (4) 300 Mg km ha 1 , LP tyres on the trail- er, and (5) 300 Mg km ha 1 , ST tyres on the trail- er.The size of the subplots was 3.2x16m.The experimental traffic was applied prior to ploughing in four successive autumns.Technical data on the vehicle and annual traffic intensity are shown in Table 2.With a traffic intens- b) LP = low-profile and ST = standard tyres on trailer.c) total weight of tractor d) not determined, but the same order as on the other years e) the soil was too wet for ST treatment Table 3. Soil moisture content at the time of experimental traffic and at field capacity (FC, b) 5,4 mm as precipitation ity of 100 Mg km ha 1 , trailer wheel tracks with low-profile tyres covered the plot area completely once.At the time of field traffic, the soil was at field capacity moisture content or wetter in the 0.2 m plough layer (Table 3).Except in the first year, the subsoil was clearly drier at 0.2-0.3m depth (80-85% of field capacity).In au- tumn 1987, the plough layer was soft and could only bear the trailer with low-profile tyres; the standard tyre plots were not loaded.These plots were loaded with a tractor just before seedbed preparation in spring 1988 (Table 2) when the topsoil (0-0.15m) was dry and the subsoil at field capacity (Table 3).The spring loading, cov- ering the plot area completely, was 53 and 159 Mg km ha' 1 in field traffic levels 3 and 5, re- spectively.
The crop was spring barley  1987) with an annual application of 110-120 kg N ha' 1 as NPK fertilizer, placed with a combi- drill.For details of the farming, see Alakukku (1996a).In these operations the axle load of machines was 0.9-2.3Mg, the tyre inflation pres- sure 100-140 kPa and the annual loading 118 Mg km ha Irrigation was carried out when the soil wa- ter content in control plots, measured by the gypsum block method at 0.15 m depth, decreased to below 50% of the plant-available water capacity (Alakukku and Elonen 1994).Plots were usually irrigated in June (Table 4).In 1987, the growing season was, however, cold and rainy and the plots were irrigated in July with rotary sprinklers.The water doses per irrigation are shown in Table 4.

Soil and plant measurements
Table 5 lists the measurements carried out during the experimental period.Soil penetrometer resistance was measured to 0.52 m depth in 0.035 m increments using a penetrometer equipped with a 12.8 mm diameter (30°angle) cone mounted on a relieved shaft and driven at approximately 0.03 m s' 1 (Andersson et al. 1980).
Results are given as the median of measuredrep- 448 AGRICULTURAL SCIENCE IN FINLAND Crop lodging 1987 all a) 1= no compaction, 5 = 300 Mg km ha 1 , standard tyres (compacted also in spring 1988) b) all tyres on a rigid surface and during the first and third loading on the field licates at each depth in a plot.The number of replicates per plot was ten except in 1987 when it was five.Soil dry bulk density was determined by the gravimetric method in 1989.Soil samples were taken from a depth of 0.03-0.48m at 0.05 m intervals with the soil core sampler de- scribed by Heinonen (1960).Three replicates taken from each depth in a plot were combined as a plot sample with a volume of 300 cm 3 .For soil moisture content during measurements and sampling see Alakukku and Elonen (1994).
The tyre/soil contact areas of the tractor-trail- er combination were determined in 1990 under soil conditions similar to those during the ex- perimental autumns.The static contact area be- tween soil and each tyre was estimated using a technique outlined by Smith and Dickson (1990).The determination was made in two positions in the field and on a rigid surface with a total of four replicates.The average ground contact pressure was calculated by dividing the wheel load by the ground contact area.
An area of 29 m 2 was harvested annually from the centre of each experimental plot.The grain yield is presented at 15% moisture content.The grain moisture content at harvest was calculated on the basis of the dry matter as determination described by Alakukku and Elonen (1995).The nitrogen content of the dry matter was measured by the near infrared reflectance technique (McGuire 1986).The area of lodged crop was evaluated as percent of the total area of a plot.
The significance of differences in soil pen- etrometer resistance, dry bulk density, grain yield and nitrogen yield data were tested with ana- lysis of variance using the General Linear Model procedure of SAS statistical programs (SAS 1990).Treatment means were tested at the 95% probability level by using the contrast statement (Milliken and Johnson 1984).If an interaction was found between irrigation and field traffic treatments, the contrast statement was performed for both irrigation levels separately.

Results
Average ground contact pressure and calculated vertical normal stress Table 6 gives the average ground contact results and the calculated average ground contact pressures.The measurements of ground contact areas were rough but the data on various tyres are comparable.The ground contact pressure was higher on the rigid surface than on the field, where tyre sinkage reduced the average ground pressure.The average rut depth with wide tyres was 0.02 m and with narrow tyres 0.11 m.It should be emphasized that stress is not uniform- ly distributed over the ground contact area.Thus, the maximum ground contact pressure at the centre of the contact area on soft soil may be as much as twice the estimated average ground pressure.On the basis of an average ground contact pressure (q) of 50 kPa and pressures listed in Table 6, the stresses to which the soil was subjected by tyres are simulated in Figure 1.The soil vertical normal stress (g z ) beneath the centre of a uniformly loaded circular ground con- tact area was calculated as a function of depth (z) using the following equation reported by Söhne (1958): Angle (a) characterizes the depth of the load- ed point below the centre of the ground contact area.Because the loaded soil was moist, a concentration factor (v) of five was used.In this calculation the soil is considered as a homogeneous, isotropic, semi-infinite, elastic medium.In the field, these conditions are seldom fulfilled, and the contact area is not circular.On the basis of this calculation, the vertical stresses due to different tyres can, however, be compared with each other.

Soil penetrometer resistance and dry bulk density
Figures 2-4 show the soil penetrometer resist- ance after the second, third and fourth loading.Figure 4 shows the soil dry bulk density after four repeated loadings.Results are given as means of irrigation levels.Irrigation did not re- duce the differences in soil properties measured in autumn between compaction levels.In 1989, the soil penetrometer resistance was, however, greater in unirrigated than in irrigated plots in the 0.18-0.38m layer since the soil was drier in unirrigated plots at a depth of 0.28-0.38m (Alakukku and Elonen 1994).
To minimize the error due to soil moisture content, an attempt was made to measure the penetrometer resistance when the soil was near field capacity (FC).In 1988, dryness (70% of FC) increased soil penetrometer resistance at 0-0.15 m depth, however, underlining the differ- ence between treatments (Figure 3).In 1989,the topsoil was damped by rain in autumn to 0.28 m depth (Alakukku and Elonen 1994).Below this depth, the soil moisture content was 90% of FC.Moreover, soil dryness increased penetrometer resistance more in plots loaded by 300 Mg km ha' 1 than in other plots in the 0.28-0.32m layer (Figure 4), thus complicating the evaluation of a cumulative increase in the intensity of com- paction from 1987 to 1989.Both soil parameters used, that is, penetro- meter resistance and dry bulk density, give less Table 6.Average ground contact pressure of the tractor-trailer combination used in experimental traffic in autumn 1990.depth beneath the centre of a circular and uniformly loaded ground contact area.
Vol. 4: 445-461.information on the effects of soil structure on drainage and crop growth than parameters sens- itive to changes in pore size distribution and the continuity of pores.Penetrometer resistance was nevertheless used because it can be determined faster than parameters requiring soil Fig. 2. Soil penetrometer resistance at field capacity as a function of depth in loaded treatments in autumn 1987.Treatment means for a given depth are followed by the same letters and those with no letters are not significantlydifferent (P BC < 0.05).
sampling and laboratory analyses.Thus, more meas- urements could be made annually.Likewise, penetrometer resistance was more sensitive to soil compaction than dry bulk density (Figure 4) as reported by Voorhees et al. (1978).For instance, in the 0.13-0.23 m layer in 1989, load- ing with 300 Mg km ha' 1 increased penetrome- ter resistance by 15% but dry bulk density by Fig. 4. Soil penetrometer resistance and dry bulk density as a function of depth in loaded treatments in autumn 1989.Soil was at field capacity (FC) to 0.28 m depth; below that at 90% of FC.Treatments and significance levels as in Fig. 2. Fig. 3. Soil penetrometer resistance as a function of depth in the control and in a loaded treatment in 1987 -1989.The loading was 300 Mg km ha 1 with trailer standard tyres in autumns 1985, 1986 and 1988, and 159 Mg km ha 1 with a tractor in spring 1988.During the measurements, soil was at field capacity (FC) except to 0.15 m depth (70% ofFC) in 1988 and below 0.28 m depth (90% of FC) in 1989.Treatment means with no P BC value are not significantly different at the 0.05 level.
Table 7. Mean relative grain and nitrogen yields as average of irrigation levels in four experimental years (control = 100).
Grain yield (%) Nitrogen yield (%) Loading (Mg km ha')  4).No statistically significant differences in dry bulk density were observed.The measured values varied greatly and three replicates in a plot were too few to allow reliable evaluation of the effects of loading on dry bulk density.7 show the grain and nitrogen yields during the experimental period.The moisture content at harvest is not reported because it was little affected by autumn corn- paction (Alakukku and Elonen 1994).This re- sult differs from those of Alakukku and Elonen (1995), who found that clay soil compaction with a 19 Mg tandem axle load reduced the moisture content when barley was harvested by acceler- ating plant ripening in compacted plots several years after compaction.

Barley yield and nitrogen uptake
in 1987, the lodging of barley and in 1988 the spring compaction of plots with standard trailer tyres complicated the interpretation of results.Soil compaction reduced crop lodging (Table 8), as Lipiec et al. (1990) and Alakukku and Elonen (1995) also observed with small grain cereals.Thus, lodging masked the negative ef- Fig. 5. Barley yield at 15% moisture content in individual years and as a mean for the four-year experimental period.
Treatment means followedby the same letters and those with no letters or P A values are not significantlydifferent (P< 0.05).fects of soil compaction on crop growth.Like- wise, irrigation in a cold and rainy year reduced crop lodging by hindering crop growth.In June, for instance, the rainfall deficiency was -21 mm (Table 4) and the mean temperature was I.7°C lower than average.In 1986In , 1988In and 1989,the rainfall deficiency in June was more than 78 mm (Table 4) and the drought at the beginning of the growing season affected crop growth in unirri- gated plots more than did soil loading in autumn (Figures 5 and 6).
In 1988, spring compaction increased the yield by 6% in plots with no irrigation (Figure 5).Barley emerged best and fastest in plots com- pacted in spring.Control plots and plots com- pacted in autumn (low-profile tyres) and spring had 355, 334 and 375 seedlings nr 2 , respectively.Tractor traffic prior to seedbed preparation sheared the aggregates near the dry soil surface.Thus, in spring-compacted plots, the amount of aggregates with a diameter smaller than 2 mm was 7 and 8 percentage units higher than in con- trol and autumn-compacted plots, respectively, in the 0-0.03 m layer.The finer seedbed in plots compacted in spring probably reduced evaporation and gave good soil-seed contact, thus improving the germination and initial growth of barley in dry, early summer (Table 4).In irrigated plots, yield was not increased by spring com- paction (Figure 5).The control plots with slow initial growth benefited from irrigation more than fast growing plots compacted in spring.

Discussion
Field traffic prior to autumn ploughing with a single-axle load of 5 Mg and a tyre inflation pressure of 150 kPa compacted the subsoil of clay loam to 0.35 m depth.When the field traffic was applied, the soil moisture content was near or higher than FC in the 0.2 m plough layer each autumn but in the subsoil only in the first year.Wetness probably made the soil more sensitive to loading.Akram and Kemper (1979) reported that the maximum compaction of mineral soils is reached when the soils are loaded at a moisture content near FC.With the same loading, wet mineral soil is compacted more and deeper than dry soil (Aura 1983, Gameda et al. 1987).
Despite ploughing to 0.2 m, frost and crop- ping, compaction due to autumn loading persisted for at least one year at 0.1-0.2m depth as shown by Voorhees et al. (1978) on clay loam.In contrast, Elonen (1980) andAura (1983) report- ed that compaction in the plough layer of clay soils due to traffic with a 3 Mg axle load prior to seedbed preparation was relieved to the next spring.As was assumed, compaction persisted in the subsoil even though the soil was cropped and the frost penetrated each year to 0.4-0.5 m depths (Alakukku and Elonen 1994).
Subsoil compaction persists for a very long time (Blake et al. 1976, Voorhees et al. 1986, Håkansson 1985).Under Finnish conditions, subsoil compaction persists for at least nine years in clay soils (Alakukku 1996 b), and increases the risk of long lasting soil degradation.The present results indicated that the load on a single axle must be less than 5 Mg under moist soil conditions to avoid soil compaction below nor- mal tillage depth.Elsewhere, axle load limits of 4 to 6 Mg on moist soils have been recomended (Danfors 1974and 1994, Voorhees and Lindström 1983, Petelkau 1984), even when the tyre inflation pressure is 50 kPa (Danfors 1994).On Finnish farms, the risk of subsoil compaction is severe during harvest, slurry spreading and till- age.With the machines used in these operations, single-axle loads of 5 Mg and a tyre inflation pressure of 100 kPa are often exceeded when the soil is moist.
No cumulative increase in the depth of soil compaction or yield losses due to annually re- Fig. 6.Nitrogen yield harvested in the grain in individual years and as a mean for the four-year experimental period.Treatment means followed by the same letters and those with no letters or P A values are not significantly different (P< 0.05).peated field traffic were found, even though the earlier compaction was not relieved before the new loading.Gameda et al. (1987) reported that the intensity and depth of clay soil compaction were increased year by year by annually repeated heavy loading on wet soil.Likewise, yield losses of spring cereals due to an annual loading of 350 Mg km ha' 1 on wet clay soils before au- tumn ploughing increased cumulatively during the first four years (Arvidsson and Håkansson 1991).In the present study, the possible cumu- lative effects of repeated loading on soil could have been evaluated better had the soil properties been measured before and after each loading.Soil penetrometer resistance was measured for the first time after two loadings.The subsoil had probably compacted primarily at the first loading because the soil was drier than FC in the 0.2-0.3m layer in the following autumns.Therefore, the effects of the later loadings were too slight to affect crop yield and to be deter- mined with the penetrometer.
Reducing trailer tyre inflation pressure from 350 to 150 kPa by increasing the tyre size did not lessen compaction in the plough layer, even though it reduced the average ground contact pressure from 162 to 82 kPa.Other workers have had contrary results (Campbell et al. 1984, Koger et al. 1985, Bolling 1987), but in their ex- periments the soil was compacted by a single tyre only.Here, on the other hand, both tractor and trailer compacted the soil, thus reducing the dif- ference between different trailer tyres.Low-profile tyres (width 500 mm) also compacted a greater area than standard tyres (width 275 mm), which may reduce the difference between tyres as found by Chamen et al. (1985).Hence the use of low-profile tyres propably did not reduce the yield or nitrogen uptake losses of barley.Cha- men et al. (1990) and Vermeulen and Perdok (1994) found that the use of a low tyre inflation pressure farming system (below 80 kPa) for four years did not affect wheat yield any differently from a high ground pressure system . By contst, Arvidsson and Håkansson (1994) found that field traffic with an intensity of 350 Mg km ha' 1 prior to autumn ploughing reduced the yield of small grain cereals by 10% when the tyre inflation pressure was 200 kPa, and by 15% when the inflation pressure was 500 kPa as a mean of four years.
The use of low-profile tyres on the trailer did not reduce the depth of soil compaction, which is consistent with the results of Danfors (1994).The trend in the present results, as in those re- ported by Alakukku and Elonen (1994) was a reduction in cumulative soil compaction in the plough sole (0.28-0.32 m), as found also by Cha- men et al. (1990).According to the theoretical calculation, the vertical normal stress below the low-profile trailer tyres was markedly less than that of standard tyres to a depth of 0.3 m.At 0.5 m depth the calculated vertical stress under both tyres exceeded, however, the limit of 30 kPa given by Grecenko (1989) for soils wetter than 70% of FC in the 0-0.3 m layer.The effect of low-profile tyres would have been greater had the ground pressure been less than the 82 kPa used.For instance, reduction of the ground con- tact pressure to 50 kPa decreased the calculated vertical stress down to 0.4 m depth consider- ably more than did the pressure of 82 kPa.
A traffic intensity of 300 Mg km ha' 1 com- pacted the soil more than that of 100 Mg km ha 1 , as reported by Koger et al. (1985) and Game- da et al. (1987).The increase in loading intens- ity did not, however, reduce either yield or nitro- gen uptake significantly.This finding is not con- sistent with many earlier yield data from experiments where the soil was ploughed after the loading (Arvidsson andHåkansson 1994, Alakukku andElonen 1995).One reason for the discrepancy results may be that, here, the increase in soil compactness was so small that it did not hinder crop growth markedly.
When the early summer was dry, irrigation increased the crop yield by 930-1990 kg ha 1 and nitrogen uptake by 8-30 kg ha 1 .
It did not, however, significantly affect the reduction in yield or nitrogen uptake due to soil compaction.By contrast, Elonen (1980) found that irrigation when the early growing season was dry reduced spring wheat yield losses due to clay soil com- paction with a 3 Mg axle load prior to seedbed Voi 4: 445-461.preparation.Likewise, Barraclough and Weir (1988) reported that irrigation reduced winter wheat yield losses due to subsoil compaction.The divergent results may be attributed to the fact that in the present study the seedbed in au- tumn-compacted plots was too coarse and hindered sprouting.Irrigation did not compensate for the effect of thin vegetation on the yield in compacted plots by increasing tillering, seed size or number of seeds per ear.
Despite compaction, the structure of this clay loam soil was good, and the soil was well drained and productive.Note also that relatively light machines were used for field work after the ex- perimental traffic.Although the soil compaction reduced yields by 80-480 kg ha' 1 (on average 5%), they were not small relative to those from many other clay soils in the surrounding area.One reason for the relatively high yields may be that the biopores and cracks present even in com- pacted soil provided pathways for root growth, despite the high penetrometer resistance of the soil.A greater relative yield reduction could be expected with soil compaction in less fertile soil, where the field work is done with heavier machines.
Soil compaction decreased nitrogen and grain yields in the same years.This is consistent with the findings of Phillips and Kirkham (1962) and Schuurman (1965) on mineral soils and small grain cereals.An equal dose of fertilizer was applied in all compaction treatments.Soil com- paction lowered the efficiency of fertilization by reducing the nitrogen uptake of crops.However, because the fertilizer was not labelled, the results do not show the amount by which com- paction reduced the uptake offertilizer nitrogen.It could be expected that, by reducing the nitro- gen uptake of crops, soil compaction would also aggravate the harmful impact ofcrop production on the environment, for instance, by increasing denitrification.
Conclusions (1) No cumulative effects due to annually repeated field traffic prior to autumn ploughing at the depth ofcompaction or yield losses were found.Except in the first year, the clay loam was drier than field capacity in the 0.2-0.3m layer, which probably impeded further subsoil compaction.
(2) The reduction in trailer tyre inflation pres- sure from 350 to 150 kPa by using low-profile tyres did not reduce compaction of the plough layer or yield reductions.It did, however, tend to reduce the intensity of compaction in the plough sole as a cumulative effect of four load- ing years.
(3) A load of 5 Mg on a single axle and a tyre inflation pressure of 150 kPa compacted a wet soil to 0.35 m depth.To avoid the risk of longlasting clay soil degradation due to subsoil com- paction, single-axle loads of less than 5 Mg should be used.
(4) Irrigation 1-2 times (18-33 mm per irrigation) early in the growing season did not significantly affect the reductions in the yield and nitrogen uptake of barley caused by soil com- paction.

Figures
Figures 5 and 6 and Table7show the grain and nitrogen yields during the experimental period.The moisture content at harvest is not reported because it was little affected by autumn corn- and soil compaction under tractor tires.Agricultural Engineering 39: 276-281, 290.Vermeulen, G.D. & Perdok, U.D. 1994.Benefits of low ground pressure tyre equipment.Developments inAgri- cultural  Engineering 11:447-478.Voorhees, W.B. & Lindström, M.J. 1983.Soil compac- tion constraints on conservation tillage in the northern corn belt.Journal of Soil and Water Conservation 38: 307-311.Nelson, W.W. & Randall, G.W. 1986.Extent and per- sistence of subsoil compaction caused by heavy axle loads.Soil Science Society of America Journal 50.G.& Nelson, W.W. 1978.Compaction and soil structure modification by wheel traffic in the northern corn belt.Soil Science Society of America Journal 42: 344-349.

Table I .
Soil particle size distribution and organic carbon content in the experimental field.

Table 4 .
\j/ m = -10 kPa).Irrigation date, water dose per irrigation and monthly rainfall deficiency in early summer during the experimental period.

Table 5 .
Measurements made during the experimental pe-

Table 8 .
Mean lodging of barley (% of plot area) three weeks before harvesting in 1987.Irrigation was carried out in July. 100 Mg km ha 1 300 Mg km ha '