ROOT ABUNDANCE OF MAIZE IN CONVENTIONALLY-TILLED AND ZERO-TILLED SOILS OF ARGENTINA ( 1 )

Maize root growth is negatively affected by compacted layers in the surface (e.g. agricultural traffic) and subsoil layers (e.g. claypans). Both kinds of soil mechanical impedances often coexist in maize fields, but the combined effects on root growth have seldom been studied. Soil physical properties and maize root abundance were determined in three different soils of the Rolling Pampa of Argentina, in conventionally-tilled (CT) and zero-tilled (ZT) fields cultivated with maize. In the soil with a light Bt horizon (loamy Typic Argiudoll, Chivilcoy site), induced plough pans were detected in CT plots at a depth of 0–0.12 m through significant increases in bulk density (1.15 to 1.27 Mg m-3) and cone (tip angle of 60 o) penetrometer resistance (7.18 to 9.37 MPa in summer from ZT to CT, respectively). This caused a reduction in maize root abundance of 40– 80 % in CT compared to ZT plots below the induced pans. Two of the studied soils had hard-structured Bt horizons (clay pans), but in only one of them (silty clay loam Abruptic Argiudoll, Villa Lía site) the expected penetrometer resistance increases (up to 9 MPa) were observed with depth. In the other clay pan soil (silty clay loam Vertic Argiudoll, Pérez Millán site), penetrometer resistance did not increase with depth but reached 14.5 MPa at 0.075 and 0.2 m depth in CT and ZT plots, respectively. However, maize root abundance was stratified in the first 0.2 m at the Villa Lía and Pérez Millán sites. There, the hard Bt horizons did not represent an absolute but a relative mechanical impedance to maize roots, by the observed root clumping through desiccation cracks.


INTRODUCTION
Deep reaching compaction in many fertile soils of the world, often managed under modern, industrialagricultural management, is an issue of great concern (Hamza & Anderson, 2005).In such soils it is often the frequent and repeated use of conventional tillage (CT: e.g.mouldboard-and disc-ploughs, disk harrows) that results in the formation of induced plough pans in subsoil (Canarache, 1991).This compacted layer can impede crop root growth when soil resistance measured with a cone (tip angle of 30 °) penetrometer exceeds the threshold of 2-3 MPa (Gupta & Allmaras, 1987;Glinski & Lipiec, 1990;Passioura, 2002).The magnitude of restriction is closely related to the soil water content, because of the resulting increases in soil resistance (Gupta & Allmaras, 1987;Raper, 2005).
For different reasons, many of these fertile agricultural soils were converted to conservation tillage systems in the last decade.Zero tillage (ZT), the paradigm of conservation agriculture, was implemented in about 95 million hectares worldwide (Lal et al., 2007).In the Pampas region of Argentina, half of the cultivated area passed to this system in the last 15 years (Díaz Zorita et al., 2002;Steinbach & Alvarez, 2006).Conventional tillage (ploughing, disking, harrowing, and so on) induces the development of so-called "plough pans" near the surface in many soils (Canarache, 1991;Taboada et al., 1998;Micucci & Taboada, 2006).These plough pans are expected to disappear under continuous ZT.In exchange, surface compaction problems have often reported in fine-textured, zero-tilled topsoils (Díaz Zorita et al., 2002;Sasal et al., 2006;Taboada et al., 1998).Although conservation tillage minimizes agricultural traffic (Hamza & Anderson, 2005;Raper, 2005), wheel tracks are not erased in ZT topsoils (Botta et al., 2006).This can contribute to increase surface compaction problems.
Among the crops used in industrial agriculture, maize (Zea mays L.) is one of the most susceptible to soil compaction (Maddonni et al., 1999;Amato & Ritchie, 2002).Maize root growth is negatively affected by subsoil compacted layers, but the consequences on maize yields are not always direct (Erbach et al., 1986;Díaz Zorita et al., 2002).Soil hardness may also have a genetic origin, as in the case of many argillic Bt horizons that behave as "clay pans" (Canarache, 1991).These hard togh B horizons may restrict the amount of available water stored in the soil profile in areas with high intra-seasonal rainfall variability (Dardanelli et al., 2003).However, clay pans are not found in all Bt horizons since the occurrence does not only depend on the clay percentage but also on the characteristics of soil prisms and the occurrence of desiccation cracks between them (Amato & Ritchie, 2002;Dardanelli et al., 2003).In Mollisols of the Pampas region of Argentina, farms converted to ZT systems coexist with farms where the CT system is still being used.Within a distance of few kilometres (Figure 1), both the topsoil and subsoil horizons consist of soils with different textural and structural properties.This situation represents an interesting opportunity to study the combined effects of man-made compaction and genetic compaction on maize root growth in the field.In this study we report field results obtained in three different soils, under CT and ZT management regimes.At each site we investigated the occurrence of man-made and genetic soil compaction and the effects on maize root abundance and root distribution.
East soils are Abruptic and Vertic Argiudolls with silty loam or silty clay loam topsoil texture and strong Bt horizon (Salazar Lea Plaza & Moscatelli, 1989).
Two agricultural fields under different managements were selected per site.Each field was divided in three adjacent plots.Soil studied managements were: (a) Long-term conventional tillage (CT): two or three passes of tandem disk harrow at 0.10 m depth and spike-tooth harrow were continuously applied over the last years.After maize sowing, weeds were mechanically controlled; (b) Zero tillage (ZT): zero till planting after long-term CT, with herbicide weed control.The number of years under continuous ZT was 9, 6, and 7 for the Chivilcoy, Villa Lía, and Pérez Millán sites, respectively.The selected agricultural fields did not differ in the technological level or crop rotation.The crop sequence over three years consisted basically of maize -wheat (Triticum aestivum L.) / soybean double cropping and full season soybean.

Determinations
In winter (June 2004), composite soil samples were taken from the 0-0.05, 0.05-0.15and 0.15-0.30m layers to determine the pH in soil suspensions (1:5 distilled water), and organic C content by the Walkley and Black method (Nelson & Sommers, 1982).
Three soil cores of 6 cm diameter from each plot were taken from the layers 0-0.05, 0.10-0.15and 0.25-0.30m to determine bulk density (BD) by the core method (Burke et al., 1986).Cores were collected between rows, unaffected by recent traffic compaction.At the sites of Chivilcoy and Villa Lía, three replicated measurements of soil penetrometer resistance (PR) were regularly made at each plot.A cone-shaped probe (tip angle of 60 °; basal diameter 1.4 cm) was driven into the soil (0.45 m deep) through consecutive falls of a 2 kg load falling from a height of 0.515 m (Burke et al., 1986).Soil PR was recorded as the number of falls required to penetrate each 5 cm of soil to a depth of 0.54 m.Apart from each penetrometer measurement, soil samples were taken from the 0-0.05, 0.1-0.15 and 0.25-0.3m layers, to determine the gravimetric water content (SWC) by oven-drying at 105 °C in the laboratory.Magnitude of soil

Description of the region
The Rolling Pampa of Argentina is one of the most important temperate crop areas of the Southern Hemisphere and covers around 5 Mha.The climate is temperate and humid, with an average annual rainfall of 940 mm, concentrated in the springsummer seasons (Hall et al., 1992), and a mean annual temperature of 17 °C (Soriano, 1991).The entire area is covered by Mollisolls, which developed from aeolian sediments (loess) under grassland vegetation (Soriano, 1991).In the West soils are mainly Typic Argiudolls with loamy texture and light Bt horizon, while in the desiccation was inferred by SWC/FC quotient (FC = field capacity).FC (g g -1 soil) was calculated using a pedo-transfer function (Campbell, 1985) validated for Pampean soils (Damiano & Taboada, 2000): FC = 0.258 -0.002 x Sand ( %) + 0.0036 x Clay ( %) + 0.03 x SOM ( %) being FC = soil water retentional -33.3 kPa matric potential; SOM = soil organic matter.
Soils were cultivated with maize from September 2004 to March 2005.Sowing dates, seed density and maize hybrids were similar at each site, for both tillage treatments.In summer (December 2004-January 2005), at maize flowering, 1 m deep soil pits were dug in each plot to determine the horizontal and vertical distribution of maize roots by a semi-quantitative method (Manichon, 1987).Each pit was 0.6 m wide with a maize plant in the centre.The abundance and vertical distribution of maize roots was determined using a 0.5 x 0.3 m rectangle subdivided in 0.05 x 0.05 m squares.In each square, root abundance was determined using a semi-quantitative scale (0 to 5).The root abundance of each 0.05 m soil layer was the average of 12 squares.Beside each pit the soil PR was evaluated in two measurements, by the same method as in the winter.On the opposite side of the pits, soil samples were taken at different depths for SWC determination in the laboratory.Magnitude of desiccation was inferred by SWC/FC quotients.

Statistical inference
At each site a factorial arrangement was applied.Factors were tillage treatment (CT and ZT) and soil depth.Differences were inferred from the analysis of variances, and in the case of significant treatment effects (p < 0.05), means were compared using contrasts (Student's test).

Soil profiles
Soils differed in the horizon sequences, topsoil texture and texture degree of Bt horizons.In Chivilcoy, the sandy loam Ap and A horizons and light textured Bt horizon do not indicate a priori the occurrence of genetic mechanical impedances in soil profile (Table 1).On the other hand, Villa Lía and Pérez Millán soils do not only have silty topsoils but also clayey and well-developed prismatic Bt horizons (Table 1).In Villa Lía the Bt horizon has no vertic features, unlike in Pérez Millán where it has slickensides and desiccation cracks.Soil organic C content was higher in the sandy loam Chivilcoy topsoil than in silty loam and silty clay loam Villa Lía and Pérez Millán topsoils (Table 2).This is likely due to different historical agricultural intensities at the study sites.In addition, organic C content was higher in ZT than in CT plots (0 to 0.15 m), at the Chivilcoy and Villa Lía sites, with modest C sequestration under zero tillage (Lal et al., 2007;Steinbach & Alvarez, 2006).Soil pHs were slightly acid in Chivilcoy and Pérez Millán, and acid in Villa Lía (Table 2), evidencing once more intense agricultural intensity in this soil.Soil pH was essentially the same in CT and ZT soils.Soil organic C losses and, to a lesser extent, soil acidification were reported as expected consequences of long-term cropping in pampas soils (Senigagliesi & Ferrari, 1993;Steinbach & Alvarez, 2006).

Soil bulk density
Soil BD was significantly affected by both soil depth and tillage, without respective interaction (Table 3).Soil depth effects were mainly caused by the lower BD determined in all cases by C enrichment in the 0-0.05 m layer.Tillage effects differed from site to site.In Chivilcoy BD was significantly lower in ZT than in CT plots, in all evaluated layers.A higher BD value was noted in the CT soil compared with ZT determined in the 0.05-0.15m layer.This BD increase may be caused by an induced plough pan at this depth (Canarache, 1991).Contrastingly, in Villa Lía and Pérez Millán soil BD was significantly higher in ZT than in CT plots, although in different soil layers (Table 3).Such increases in BD can be the result of surface compaction under zero tillage (Thomas et al., 1996).

Soil water content
In the winter, SWC was determined in the top 0.3 m at the sites Chivilcoy and Villa Lía (Table 4a,b).The SWC/FC quotient indicated the magnitude of soil desiccation on each sampling date.In the winter, soils reached a field capacity of 0.73-0.95,with significantly higher SWC in ZT than in CT plots in topsoil in Chivilcoy.In Villa Lia, SWC/FC quotients were lower and only differed between management in the top 5 cm (CT > ZT).In the summer, only the Villa Lía soil showed significant differences between tillage treatments.Even in summer and under maize crop flowering, the magnitude of soil desiccation was not high (i.e.high SWC/FC quotients) in the deep profile of Chivilcoy (0.95-1 m layer) and in the intermediate profile (0.45-0.5 m layer) of Villa Lía and Pérez Millán sites (Table 4).In the latter, high SWC were due to water storage in the clayey Bt horizons at  (Amato and Ritchie, 2000).

Soil penetrometer resistance
Soil resistance is highly dependent on soil water condition (Gupta & Allmaras, 1987); lower PR values were thus observed in winter than in summer (Figure 2).Soil PR was significantly affected by a tillage-by-depth interaction in the winter.PR values increased from 1-3 MPa in topsoil to 5-7 MPa in subsoil (0.45 m) in the Chivilcoy and Villa Lía soils.Taking into account the wet soil conditions in these periods (Table 4), the higher PR values were the result of profile anisotropy in deep Bt horizons.Statistical differences observed between treatments at two soil depths were little relevant in magnitude.
In summer, soil PR was affected by significant and independent effects of depth and tillage in Villa Lía  2).Depth effects differed from site to site.In Chivilcoy, soil PR was significantly lower in topsoil than in the rest of the profile.Despite a higher water content under CT than ZT, PR was higher at 0.15 m.Taking into account the higher BD determined in CT soil at the same depth (Table 3), this PR increase can be ascribed to an induced plough pan by long-term conventional tillage (Canarache, 1991).It is likely that after nine years ZT this induced compaction is recovered at the Chivilcoy site.Similar recoveries were also found in other medium-textured soils of the region (Taboada et al., 1998;Micucci & Taboada, 2006).Soil PR was as high as 15 MPa in the 0.1-0.2m layer of the fine textured Pérez Millán soil.This PR increase occurred both in CT and ZT soils, so that it is questionable to ascribe it to an induced plough pan.No clear explanation was found for the persistence in ZT plots either, since the high number of years (seven) would have allowed the recovery of porosity in the untilled soil (Rhoton, 2000).Therefore, the PR data recorded here could be ascribed to a genetic cause.At the same depth, there is a thin BA horizon with a lower organic C content than in the higher A horizon and with a lower clay content than in the deeper Bt horizon (Tables 1 and 2).Considering the very low GWC found in summer at 0.2-0.25 m depth in both CT and ZT plots of Pérez Millán soil (Table 4), it can then concluded that this BA horizon behaved as a natural compacted layer in the summer.Soil PR decreased sharply below 0.2 m, despite the strong Bt horizon existing at this depth (Table 1).Unlike the desiccated upper horizons, higher GWC was determined at this depth (Table 4).The combined action of water and clay plasticity may have allowed an easy penetration of the probe in this clayey subsoil (Pilatti & Orellana, 2000).
Although the Villa Lía soil also has a silty clay loam topsoil and clayey subsoil, soil PR behaved completely different here (Figure 2d).Soil PR was not only significantly higher in CT than in ZT plots, but also increased significantly with depth.This PR increase was not related to the higher GWC found in this Bt horizon (Table 4), and can be ascribed to the low cracking potential of this subsoil.In contrast, soil cracking is very high in the vertic Pérez Millán subsoil, which could help to decrease soil hardness.Both Bt horizons of Villa Lía and Pérez Millán soils differed substantially from each other in clay content (Table 1) and have different clay mineralogy.The expansible clay mineralogy of Pérez Millán soil determined lower mechanical impedance, incompatible with a clay pan.This different behaviour can be ascribed to the characteristics of soil prisms and the occurrence of desiccation cracks among them (Amato & Ritchie, 2002;Dardanelli et al., 2003).
A threshold of 2 MPa, measured with probes (tip angle of 30 °), has often been mentioned to detect mechanical impedances for crop root growth (Gupta and Allmaras, 1987;Glinski & Lipiec, 1990;Passioura, 2002).However, soil resistance can increase by 68 % when 60 º probes are used instead of 30 ° probes (Voorhees et al., 1975).Soil PR values determined here exceeded both threshold limits by far.In Argentina, Pilatti & Orellana (2000) observed that a threshold of 6 MPa is more suitable for Pampean soils similar to those studied here, using a 60 ° probe.

Maize root abundance
The abundance of maize roots was affected by independent and highly significant effects of tillage and depth (Figure 3a,b,c).Tillage effects were most evident in Chivilcoy, where maize root abundance was significantly higher in ZT than in CT plots below 0.1 m.This can be explained by the higher BDs (Table 3) and PRs (Figure 2c) observed in the CT plot, due to the occurrence of induced plough pans.This situation represents an effect known as "shadow effect", as shown by Tardieu (1988) for surfacecompacted soil layers.Even when there is no mechanical impedance in the rest of the profile, the occurrence of surface compaction affects root exploration in the zone of the profile below this compacted layer.Map distributions of maize roots showed that a major soil volume of the CT plot has little or no roots below 0.75 m (Figure 4a).In the ZT plot of Chivilcoy Maize root exploration occurs throughout the profile (Figure 4b).A differentiated water and nutrient absorption by maize can therefore be expected in each tillage treatment (Tardieu, 1988;Passioura, 2002).
Although soil profiles differed greatly in summer in Villa Lía and Pérez Millán sites (Figure 2d,e), the root abundance profiles did not (Figure 3b,c).Tillage affected maize root abundance in both soils significantly, while the magnitude of this effect was generally little relevant.An exception was the higher maize root abundance under ZT in Pérez Millán profile below 0.4 m (Figure 3c).Because of this root abatement only 10 % relative maize abundance was observed below 0.8-0.9 m.However, taking the sharp PR drop with depth in Pérez Millán into account, maize root abundance decreases with depth were due to the sharp PR increase detected in the deeper BA horizon under both CT and ZT.

Figure 1 .
Figure 1.Geographical location of studied sites in the province of Buenos Aires, Argentina.

Figure 2 .
Figure 2. Soil penetrometer resistances in different layers of conventionally-tilled (CT) and zero-tilled (ZT) soils in winter and summer.Standard errors of means are indicated by bars.ANOVA tables are also included.(*) indicates statistical differences between management at a specific soil depth.

Figure 3 .
Figure 3. Relative abundance of maize roots in 1 m deep soil profiles in conventionally-tilled (CT) and zero-tilled (ZT) soil profiles of study sites.Standard errors of the means are indicated by bars.ANOVA tables are also included.

Figure 4 .
Figure 4. Spatial distribution of maize roots in 0.6 wide and 1 m deep soil profiles under conventional tillage (CT) and b) zero tillage (ZT) at the studied sites.

Table 1 . Soil properties of the studied sites
Data were abtained from published soil maps(INTA, 1980a, b)and checked in situ.

Table 4 . Soil water retention at -33.3 kPa matric potential (field capacity = FC), gravimetric soil water contents (SWC) in winter and in summer, and the FC/SWC quotients only
(Figure