Root hairs improve root penetration, root–soil contact, and phosphorus acquisition in soils of different strength

Abstract

Root hairs are a key trait for improving the acquisition of phosphorus (P) by plants. However, it is not known whether root hairs provide significant advantage for plant growth under combined soil stresses, particularly under conditions that are known to restrict root hair initiation or elongation (e.g. compacted or high-strength soils). To investigate this, the root growth and P uptake of root hair genotypes of barley, Hordeum vulgare L. (i.e. genotypes with and without root hairs), were assessed under combinations of P deficiency and high soil strength. Genotypes with root hairs were found to have an advantage for root penetration into high-strength layers relative to root hairless genotypes. In P-deficient soils, despite a 20% reduction in root hair length under high-strength conditions, genotypes with root hairs were also found to have an advantage for P uptake. However, in fertilized soils, root hairs conferred an advantage for P uptake in low-strength soil but not in high-strength soil. Improved root–soil contact, coupled with an increased supply of P to the root, may decrease the value of root hairs for P acquisition in high-strength, high-P soils. Nevertheless, this work demonstrates that root hairs are a valuable trait for plant growth and nutrient acquisition under combined soil stresses. Selecting plants with superior root hair traits is important for improving P uptake efficiency and hence the sustainability of agricultural systems.

Barley, cereals, drought, Hordeum vulgare L., mechanical impedance, plough pan, rhizosheath, rhizosphere, tillage.

Introduction

Although limited phosphorus (P) availability and high soil strength are known to be major constraints to plant growth, few studies have examined the effect of these stresses in combination. This is particularly relevant in the selection of root traits for improved P uptake, as it is not clear how they will perform under high-strength conditions when root growth is impeded. The extent of mechanical impedance affecting root growth should not be underestimated, as many soils under mechanized agriculture can rapidly increase in strength upon drying (Bengough, 1997; Bengough et al., 2011), with the effects exacerbated by traffic (Newton et al., 2012) or inherent mechanical properties (Whitmore and Whalley, 2009). This is a typical environment in some important cropping areas, such as Western Australia and in the highest yielding areas for cereals globally in northern maritime Europe (Valentine et al., 2012). Moreover, almost all farming regions have seen a dramatic shift in soil cultivation as well, with non-inversion tillage to <10cm commonplace and >20% of the USA and almost half of Argentinian agricultural land under zero-tillage (Farage et al., 2007).

Soils with low P availability are widespread and are a significant constraint to global agricultural productivity (Lynch, 2007). P fertilizer is used to supply P; however, there is a need to manage inputs of P fertilizer in both an economically and an environmentally sustainable manner (White et al., 2012). P fertilizers are largely derived from finite reserves of rock phosphate and are, or will become, an increasingly expensive and scarce input (Dawson and Hilton, 2011). Mismanagement of P fertilizers also poses a risk for eutrophication of the wider environment (Cordell et al., 2009). Improving the ability of plant root systems to forage for this relatively immobile nutrient is a key strategy for improving P efficiency in agricultural systems (Richardson et al., 2011).

Improving P uptake efficiency could be achieved through a range of root morphological and architectural adaptations (White et al., 2013). Plants increase total soil exploration by increasing root length, increasing root branching, increasing specific root length, and modifying branching angle (Lynch and Brown, 2001; Gahoonia and Nielsen, 2004a; Lynch, 2007). For many plant species, root hairs are a particularly important trait for P acquisition (Gahoonia and Nielsen, 1998; Gahoonia et al., 2001). These single-cell extensions on the surface of roots increase the surface area of the root and therefore increase root–soil contact, and can contribute up to 80% of P uptake (Jungk, 2001). Root hairs are also thought to be useful for plants to acquire resources across air-filled voids and aid penetration through strong soils by acting as bracing structures (White and Kirkegaard, 2010; Bengough et al., 2011).

Significant inter- and intra-specific variation exists for root hair traits, and this has been linked to P uptake. Plants with longer, denser root hairs exhibit greater P uptake and plant growth in P-deficient soils (Gahoonia and Nielsen, 1997, 2004a; Brown et al., 2012). Genetic variation for root hair traits, particularly root hair length, can be exploited in breeding for improved P uptake efficiency and P fertilizer use efficiency in crops (Brown et al., 2013b). However, root hairs are highly responsive to environmental conditions. For example, root hair initiation and elongation are restricted in high-strength and aluminium-toxic soils (Cornish et al., 1984; Hoffmann and Jungk, 1995; Haling et al., 2010, 2011).

When the combined environmental pressures summarized above are integrated, the possible challenge to P acquisition is enormous. Changes in tillage practice, for instance, can decrease the depth at which impeded layers are encountered, but modern varieties tend to be selected and trialled on ploughed soils atypical of emerging commercial practices. With changes to tillage, root proliferation will be altered because of differences in the depth distribution of soil mechanical properties, measured typically as penetrometer resistance. At penetrometer resistances >1MPa, root elongation is restricted (Passioura, 2002). This restricted root growth limits the ability of plants to explore the soil for water and nutrients. This is particularly important for P acquisition, for which plants rely upon maximizing the volume of soil explored by roots (mainly topsoil) to access this immobile nutrient (Lynch, 2007). However, studies of P uptake in high-strength soil have demonstrated that improved root–soil contact increases P uptake per unit root length (Cornish et al., 1984; Shierlaw and Alston, 1984). This might have profound effects on the ability of plants to acquire P under different tillage regimes and could lead to quite different physiological and genetic responses to P deficiency under different soil conditions. For example, traits and chromosomal regions responsible for P acquisition can differ between minimum tillage and conventional systems (George et al., 2011). Therefore, the interaction between soil strength and P acquisition is worth pursuing.

It is not known whether improved root–soil contact in high-strength soil can over-ride the benefit of root hairs for P uptake. While it is hypothesized that the value of root hair traits for P uptake may be diminished in high-strength soils (due to shorter root hairs), they might still be a valuable trait for anchoring roots and aiding root penetration into and through high-strength soil (Czarnes et al., 1999; Bengough et al., 2011). In this study, a range of root hair genotypes of barley (Hordeum vulgare L.) are used to investigate the value of root hairs for root penetration, root–soil contact, and P uptake in high-strength soils. To date, it is not known if plant lines of a particular species with differences in root hair traits will exhibit these phenotypic differences under conditions that are known to affect their growth. If the genotype is not robust to a range of environmental conditions, identifying improved genotypes for P acquisition will be complicated and the postulated benefits possibly constrained (Brown et al., 2013b).

Materials and methods

Three separate experiments were conducted. Experiment 1 examined how root hairs impacted root penetration into high-strength layers, using controlled conditions of soil packing and growth. In Experiment 2, the effect of high-strength soil on root proliferation and P acquisition was examined, again using controlled laboratory packing of sieved soil and growth in a glasshouse. The experiments are then validated in field cores in Experiment 3, where differences in soil mechanical properties with depth are assessed using soils that have been subjected to different depths of cultivation for eight continuous years (Newton et al., 2012).

For all experiments, a sandy loam soil was used that was collected from the 0–10cm layer of a Cambisol (FAO, 1998) near Dundee, Scotland as described by Brown et al. (2012). Although this soil has a relatively high Olsen P of 54.1mg P kg^–1, barley has previously been shown to be P responsive when grown in this soil (George et al., 2011; Brown et al., 2012).

Experiment 1: impact of root hairs for root penetration into high-strength layers

This experiment investigated the ability of two different root hair genotypes of barley, H. vulgare L. (i.e. genotypes with or without root hairs), to penetrate from a low-strength into a high-strength soil.

Soil and experimental conditions

Soil (<5mm) was wet to 0.133g g^–1 with deionized water and allowed to equilibrate overnight at room temperature. The soil was then sieved to <2mm and packed into cylindrical PVC pots (63mm internal diameter; 100mm height; mesh base) in two layers to form a low-strength upper layer over a high-strength lower layer. Soil strength was measured using a mechanical test frame (0.97mm diameter 30 ° cone-angle penetrometer with 80% relieved shaft; 15mm extension; 50 N load cell; Model 5544, Instron, High Wycombe, UK) and soil matric potential measured using an SWT-5 miniature soil tensiometer (Delta-T Devices, UK). The lower layer had three different strength treatments to reflect high [1.6g cm^–3; penetration resistance of 3.05±0.18MPa (SE), n = 3] and very high (1.7g cm^–3; penetration resistance of 4.45±0.33MPa) soil strengths, plus a low-strength control that was also used for the upper layer (1.2g cm^–3; penetration resistance of 0.03±0.01MPa). Soil matric potentials were all within the readily available range; –40 kPa (±8 SE, n=3; air-filled porosity of 0.39 m³ m^–3), –43.6 kPa (±1.1 SE, n=3; air-filled porosity of 0.18 m³ m^–3), and –44.5 kPa (±0.2 SE, n=3 air-filled porosity of 0.13 m³ m^–3) for the 1.2, 1.6, and 1.7g cm^–3 treatments, respectively. For the lower layer, 224, 299, and 318g dry soil equivalent was weighed into the pots (1.2, 1.6, and 1.7g cm^–3, respectively), tamped against a hard surface, and then pressed to a height of 60mm using a hydraulic press. For the upper layer, 112g dry soil equivalent was weighed on top of the lower layer and gently tamped to a height of 30mm. As the stress exerted on the upper layer of soil was much smaller than that on the lower layer, mechanical deformation will have been isolated to the lower layer alone. The water content was selected so that the air-filled porosity of the most compact soil treatment, 1.7g cm^–3, was slightly greater than the critical cut-off of 0.10 m³ m^–3 where hypoxia limits root function.

One of three lines of root hairless (NRH, i.e. no root hairs) and root hair (RH) genotypes of barley (Brown et al., 2012) were selected for the study. While these lines were not genetic isolines, and were likely to have multiple mutations, these genotypes were selected as they had been demonstrated to be robust over numerous studies (Brown et al., 2012, 2013a). The RH genotype refers to one of the long root hair (i.e. LRH) genotypes originally described by Brown et al. (2012); in this case, RH was considered a better descriptor than LRH given that the root hair length of this genotype did not differ significantly from that of the wild type.

Seeds were germinated overnight on 0.5% distilled water agar [1g of BDH Agar (VWR, UK) per 200ml of distilled water]. Three seeds were sown in each pot in a triangular pattern (10mm spacing; 10mm depth) with radicles directed towards the centre of the pot. Plants were grown in a glasshouse (18/14 ^○C day/night temperature) with ~16h of daylight supplemented with artificial lighting to maintain a minimum light intensity of 200 μmol quanta m^–2 s^–1. Plants were watered daily by adjusting pot mass to weight with deionized water to return to an initial condition of 0.133g g^–1. The design was a randomized block design with four replicates.

Harvest and measurements

Plants were harvested 4 d after sowing. For the low-strength lower layer treatment (1.2g cm^–3), root systems were removed from pots, soil washed from the roots, and the number and length of each seminal root length counted, or measured with a ruler. For the high-strength lower layer treatments (1.6g cm^–3 and 1.7g cm^–3), the upper layer was removed from the pot by up-ending the pot and gently tapping out the soil. The number of seminal roots penetrating the high-strength layer was determined by pulling each root gently with tweezers to establish if the root was anchored in the high-strength layer. It was generally evident if the roots were anchored in the high-strength layer as the roots remained attached to the high-strength layer upon up-ending the pot. Roots were then washed from the high-strength layer. As per the control soil, the number and length of each seminal root were recorded. In addition, the portion of each root that was growing in the high-strength layer was determined by measuring the length of the root that was thickened and had greater discoloration due to soil adhering to the root. This was considered a reliable indicator of root growth in the high-strength layer as root diameter increases in response to high-strength soil (e.g. Haling et al., 2011), and it has also been observed that it is more difficult to remove soil from roots that have been growing in high-strength soil, including root hairless genotypes that tend not to form large rhizosheaths (Brown et al., 2012).

Experiment 2: value of root hairs for P uptake in high-strength soil

The effect of soil strength on the shoot growth, root growth, and P uptake of the two root hair genotypes used in Experiment 1 was investigated in a glasshouse experiment.

Soil and experimental conditions

Soil additional to that used in Experiment 1 was collected and sieved to <5mm. Two P treatments were applied to the soil; an unamended treatment (P0) and a P-sufficient rate of 500mg P kg^–1 (P500) that was based on the data of George et al. (2011). For each P treatment, 40kg dry soil equivalent was mixed in a cement mixer for 20min. For the P500 treatment, 87.89g of KH₂PO₄ was added to the soil prior to mixing. Soil was incubated at 14–18 ^○C for 23 d at 0.133g g^–1 water content, as per Experiment 1. After incubation, soil water content was adjusted to 0.133g g^–1.

Soil of each P treatment was packed into separate cylindrical PVC pots (50mm internal diameter; 200mm height; mesh base) to a height of 180mm. Pots were packed with 424, 566, and 601g of dry soil equivalent to achieve bulk density treatments of 1.2, 1.6, and 1.7g cm^–3, respectively. Soil was added to pots in two equal batches; soil was tamped and pressed to the appropriate height as per Experiment 1. An additional three pots of each soil bulk density treatment were prepared, and soil strength and matric potential measured as described in Experiment 1. Pre-germinated seeds of the NRH and RH barley genotypes were sown 10mm deep into pots of each treatment, with one seed per pot. The design was a randomized block design with five replicates and two harvest times. Plants were grown in a glasshouse under the same conditions as Experiment 1 and pots watered daily to mass. At 13 d and 20 d after sowing, plants for the second harvest were watered with 10ml per pot of P-free nutrient solution [25mM (NH₄)₂SO₄, 2mM KNO₃, 1mM MgSO₄, 10mM Ca(NO₃)₂, 80 μM FeEDTA, 30nM H₃BO₃, 6 μM CuSO₄, 6 μM MnSO₄, 0.6 μM ZnSO₄, 42nM NH₄Mo₇, 12 μM Co₄(NO₃)₂].

Harvest and measurements

Plants were harvested at 11 d and 21 d after sowing. At the first harvest, roots were recovered from the soil with rhizosheaths intact. Root systems with the rhizosheath still attached were weighed and the soil was washed from the root. At the second harvest, roots were washed directly from the soil in each pot. Roots were stored at 4 ^○C in 50% ethanol. Roots were scanned [400 dpi; Epson Expression 1640xL flatbed scanner (Epson UK, London, UK)] and root length and average root diameter analysed using WinRHIZO (Regent Instruments, Quebec, Canada). For the 11-day-old root systems, root systems were placed in a layer of water and two seminal roots on each plant were photographed 40–60mm from the root tip using a Leica MZFIII microscope fitted with a Leica DC480 camera with a Leica 10446561×0.63 lens [Leica Microsystems (UK) Ltd, Milton Keynes, UK]. Twelve root hairs (six per root) were randomly selected and their length measured using ImageJ (US National Institutes of Health, Bethesda, MD, USA).

Shoots were cut at the seed and oven-dried for 72h at 70 ^○C. Whole shoot samples were weighed to determine shoot dry mass and then milled to a powder. Subsamples of 50mg were digested for 20min at 180 °C in 3ml of 15.8M HNO₃ (Aristar grade, VWR International, Poole, UK), followed by oxidation for 20min at 180 °C with 1ml of H₂O₂ in closed vessels using a MARSXpress microwave oven (CEM, Buckingham, UK). Digested samples were diluted to a final volume of 50ml with de-ionized water and the concentrations of P in diluted digests were determined by reaction with malachite green (Irving and Mclaughlin, 1990).

Experiment 3: response of root hair genotypes to minimum and conventional tillage treatments in intact cores

The impact of root hairs on the ability to acquire P from soil subjected to different tillage treatments, and therefore different soil physical conditions, was tested by growing a number of root hairless and haired genotypes along with the wild-type in intact soil cores from an 8-year-old tillage trial.

In December 2011, when soils were at approximately FC (field capacity), soil cores (height 130mm; internal diameter 100mm; and average soil mass of 2.7kg) were taken from a tillage trial at a field site located near Dundee, Scotland (George et al., 2011; Newton et al., 2012). A range of cultivation treatments were established in triplicate in autumn 2003 that imposed different levels of soil disturbance. This study used samples taken from the conventional plough treatment and the minimum tillage treatment, which represented the two most commonly used cultivation methods. It was known that these two treatments created a plough plan beneath cultivation depth that has increased soil strength at 25cm and 7cm depth, respectively (George et al. 2011). For the previous eight cropping seasons, each experimental plot was split in two, with half sown as winter barley and the other half as spring barley. Cores were taken in groups of three from the winter barley half, just prior to soils being excavated for other purposes. The winter trials were combine-drilled with a compound fertilizer at a rate of 70kg P and 105kg K ha^–1, and had received applications of 51kg N ha^–1 and 69kg N ha^–1 at growth stage (GS) 22–24 and GS 31–32, respectively. The spring trials were combine-drilled with a compound fertilizer at a rate of 77kg N, 14kg P, and 49kg K ha^–1, and had received an additional 33kg N, 6kg P, and 21kg K ha^–1 at GS 22–24.

Soil cores were stored outside for 2 weeks before being brought into the glasshouse. The glasshouse was warmed to 18/14 ^○C (day/night) with ~16h of daylight at a minimum light intensity of 200 μmol quanta m^–2 s^–1 ensured by supplementary lighting. Over the next 2 weeks, weeds were allowed to germinate (and were removed from the pots) and water content was kept constant by mass. Once weeds had stopped emerging, six genotypes exhibiting variation in root hair length [three NRH and three RH, including those used in Experiments 1 and 2; originally referred to by Brown et al. (2012) as LRH] were selected from the mutant population, plus a wild-type control, and planted in five replicates into the soil cores. Seeds of uniform size were selected and germinated on 0.5% distilled water agar [1g of BDH Agar (VWR, UK) per 200ml of distilled water] prior to planting, until their radicles were between 5mm and 10mm long. Each pot was sown with three germinated seeds of one of the six mutant genotypes or the wild type. The soils were maintained at ~80% FC during the growth period by watering to mass with distilled water daily, and all required nutrients, except P, were provided weekly by addition of 25ml per pot of a nutrient solution as described in Experiment 2. This ensured plant growth was not limited by nutrients other than P. Plants were grown in a randomized design and pots were rotated between glasshouse benches regularly to minimize effects of possible environmental gradients.

After 8 weeks, once the plants had reached GS 39–59 (dependent on the treatment), they were harvested by cutting the stem at the soil surface. Total shoot biomass was measured after oven drying at 60 °C for 5 d. Shoot P concentration was measured in powdered flag leaf samples (50mg) as per Experiment 2.

Data and statistical analysis

Data were checked for normal distribution and log-transformed [i.e. log₁₀(x+1)] where appropriate, with analysis by GenStat 15th Edition (VSN International, UK) using analysis of variance (ANOVA).

For Experiment 1, measurements of the number and length of seminal roots were averaged for each plant, and a mean determined for each pot. The number of roots penetrating into the lower layer was expressed as a percentage of the total number of seminal roots per pot that were considered long enough (i.e. >30mm) to be growing in the lower layer. Data were analysed using general ANOVA, with bulk density of the lower layer and genotype as factors.

For Experiment 2, plants from the first harvest were used for measurements of root hair length and rhizosheath. Plants from the second harvest were used to calculate parameters for P uptake. Total shoot P was calculated as the product of shoot dry mass and shoot P concentration. P uptake per unit root length was calculated by dividing total shoot P by the total root length. Data were analysed using general ANOVA, with bulk density, P, and genotype as factors.

For Experiment 3, total shoot P was calculated as described for Experiment 2. For presentation purposes, the three genotypes representing each phenotype (NRH and RH) were averaged, meaning that each mean data point consists of 15 replicates. Data were analysed using general ANOVA, with genotype and tillage as factors.

It should be noted that in Experiment 2 very short root hairs were observed on the NRH genotype. For consistency, this genotype is continued to be referred to as root hairless, i.e. NRH.

Results

Experiment 1: impact of root hairs on root penetration into high-strength layers

Mean values of the root growth traits for two root hair genotypes (NRH and RH) for the different soil density treatments are presented in Table 1. In the control treatment (1.2g cm^–3 lower layer), 100% of the seminal roots of both genotypes were found growing in the low-strength lower layer, and the length of these roots did not differ between genotypes (P > 0.05). As the bulk density of the lower layer increased to 1.6g cm^–3, the average seminal root length of the two genotypes decreased by up to 50% of that measured in the loose soil (P < 0.05). The ability of the NRH genotype to penetrate into, and grow through, the high-strength lower layer was also restricted. At 1.7g cm^–3, only 1% of roots of the NRH genotype were found growing in the lower layer, compared with 88% of the RH genotype. Furthermore, in the high-strength layer, the length of the roots of the RH genotype were, on average, more than double the length of those of the NRH genotype (P < 0.05). The two genotypes did not differ in the total number of seminal roots per plant.

Table 1.

Root growth attributes of a root hairless and root hair genotype of barley grown for 4 d in pots with a low bulk density upper layer (1.2g cm^–3; 30mm depth) over three different lower layers (1.2, 1.6, and 1.7g cm^–3; 60mm depth)

	Seminal roots penetrating into lower layer (%)		Seminal root length growing in whole pot (cm root–1)		Seminal root length growing in high-strength lower layer (cm root–1)		No. of seminal roots (number plant–1)
	No root hairs	Root hairs	No root hairs	Root hairs	No root hairs	Root hairs	No root hairs	Root hairs
Bulk density of lower layer
1.2g cm–3	100 a	100 a	9.01	8.74	–	–	7	7
1.6g cm–3	80 a	96 a	4.81	5.54	0.92	1.61	7	6
1.7g cm–3	1 b	88 a	4.30	4.92	0.35	1.64	7	7
ANOVA (P < 0.05)	LSD	P-value	LSD	P-value	LSD	P-value	LSD	P-value
Bulk density (BD)	*	<0.001	0.52	<0.001	–	0.35	–	0.13
Genotype	*	<0.001	–	0.09	0.62	0.005	–	0.055
BD×genotype	*	<0.001	–	0.11	–	0.29	–	0.34

*Data were log transformed.

Different letters denote significant differences between treatments.

Experiments 2 and 3: value of root hairs for P uptake in high-strength soil, and response of root hair genotypes to minimum and conventional tillage treatments in intact cores

Mean values of root and shoot parameters for two root hair genotypes (NRH and RH) grown in soil at three bulk densities (1.2, 1.6, and 1.7g cm^–3) and at two rates of P (P0 and P500) are presented in Table 2, Table 3, and Fig. 1. Overall, root dry mass, root length, shoot dry mass, tiller count, and P uptake parameters (i.e. shoot P concentration, total shoot P, and P uptake per unit root length) of both genotypes responded positively to the application of P. Increases in soil bulk density were found to restrict root length, and, depending upon genotype and P treatment, more often had neutral or positive effects on shoot dry mass and P uptake parameters (Fig. 1).

Fig. 1.

(a) Total shoot P (mg) and (b) P uptake per unit root length (mg m^–1) of a root hairless (NRH) and root hair (RH) genotype of barley grown in unamended (P0) and P-amended (P500) soil at bulk densities of 1.2, 1.6, and 1.7g cm^–3. Bars show the least significant difference (P < 0.05; n=5) for the interaction between P, genotype, and bulk density.

Table 2.

Effect of soil bulk density and phosphorus amendment (0 or 500mg P kg^–1) on the root dry mass (mg), total root length (cm), average root diameter (mm), root hair length (mm), and rhizosheath (mg dry soil cm^–1 root length) of a root hairless and a root hair genotype of barley grown for 11 d (root hair length and rhizosheath) or 21 d (root dry mass, total root length, and average root diameter)

Phosphorus rate	Root dry mass (mg)				Total root length (cm)				Average root diameter (mm)				Root hair length (mm)				Rhizosheath (mg cm–1)
	No root hairs		Root hairs		No root hairs		Root hairs		No root hairs		Root hairs		No root hairs		Root hairs		No root hairs		Root hairs
	P0	P500	P0	P500	P0	P500	P0	P500	P0	P500	P0	P500	P0	P500	P0	P500	P0	P500	P0	P500
Bulk density of soil
1.2g cm–3	51	134	133	216	762	1252	1191	1234	0.40	0.47	0.63	0.65	0.04	0.02	0.89	0.77	0.8 a	0.3 a	32.7 d	23.3 e
1.6g cm–3	76	163	80	155	975	1282	959	1220	0.45	0.51	0.69	0.58	0.05	0.05	0.73	0.57	1.7 a,b	2.9 b	23.8 d,e	18.3 e
1.7g cm–3	69	143	103	133	624	1080	832	952	0.52	0.55	0.65	0.61	0.11	0.08	0.74	0.54	5.9 c	2.5 b	17.4 e	18.9 e
ANOVA	LSD		P-value		LSD		P-value		LSD		P-value		LSD		P-value		LSD		P-value
Bulk density (BD)	15*		0.056		106		<0.001		0.03		0.011		–		0.118		0.080**		0.001
Genotype	15		<0.001		72*		0.099		0.02		<0.001		0.0135**		<0.001		0.065**		<0.001
Phosphorus (P)	15		<0.001		86		<0.001		–		0.690		0.0135**		<0.001		0.065**		0.001
BD×genotype	26		<0.001		125*		0.055		0.04		0.002		0.0233**		<0.001		0.113**		<0.001
BD×P	–		0.185		–		0.971		0.04*		0.081		–		0.354		–		0.353
Genotype×P	–		0.298		122		0.003		0.03		<0.001		0.0190**		0.030		–		0.373
BD×P×genotype	–		0.359		–		0.181		0.05*		0.088		–		0.885		0.160**		0.027

P < 0.05; or *indicates LSD value for P < 0.10; **indicates LSD for log-transformed.

Different letters denote significant differences for log-transformed data.

Table 3.

Effect of soil bulk density and phosphorus amendment (0 or 500mg P kg^–1) on the shoot dry mass (mg plant^–1), shoot P concentration (mg g^–1), and number of tillers (number plant^–1) of a root hairless and root hair genotype of barley grown for 21 d

Phosphorus rate	Shoot dry mass (mg plant–1)				Shoot P concentration (mg g–1)				No. of tillers (no. per plant)
	No root hairs		Root hairs		No root hairs		Root hairs		No root hairs		Root hairs
	P0	P500	P0	P500	P0	P500	P0	P500	P0	P500	P0	P500
Bulk density of soil
1.2g cm–3	0.07	0.24	0.24	0.33	1.31	5.07	4.02	6.90	1.0	3.5	2.6	3.3
1.6g cm–3	0.12	0.37	0.28	0.36	1.28	7.97	3.84	7.94	1.0	4.3	3.2	3.6
1.7g cm–3	0.13	0.37	0.24	0.34	1.04	7.89	3.40	8.61	1.0	4.4	2.2	4.0
ANOVA (P < 0.05)	LSD		P-value		LSD		P-value		LSD		P-value
Bulk density (BD)	0.03		<0.001		0.24		<0.001		–		0.267
Genotype	0.03		<0.001		0.20		<0.001		0.4		0.004
Phosphorus (P)	0.03		<0.001		0.20		<0.001		0.4		<0.001
BD×genotype	0.04		0.019		0.34		<0.001		–		0.720
BD×P	–		0.276		0.34		<0.001		–		0.120
Genotype×P	0.04		<0.001		0.28		<0.001		0.6		<0.001
BD×P×genotype	–		0.270		0.48		<0.001		–		0.448

Root and shoot growth

On average, the RH genotype developed larger root systems and had greater shoot development and P uptake than the NRH genotype (P < 0.05). However, this varied depending upon both the P treatment and the bulk density of the soil. In the P0 soils, shoot dry mass, tiller count, and P uptake parameters of the RH genotype were between 2- and 4-fold larger than those of the NRH genotype (P < 0.05). However, in the P500 soils, performance of the NRH genotype increased relative to that of the RH genotype, and in most cases the genotypes did not significantly differ in total root length, shoot dry mass, tiller count, or P uptake. Similarly, at lower soil bulk densities (1.2g cm^–3) and hence softer soil, the root dry mass, total root length (P < 0.10), shoot dry mass (1.2g cm^–3 and 1.6g cm^–3; P < 0.05), and P uptake (total shoot P and shoot P concentration; 1.2g cm^–3; P < 0.05) of the RH genotype were up to 5-fold larger than those of the NRH genotype. In the high P soil, as the bulk density of the soil increased, the growth of the NRH genotype tended to increase relative to that of the RH genotype, with differences between the genotypes not detected. This was also apparent when plants were grown in intact cores that were taken from an 8-year-old field-based tillage trial (Fig. 2). The NRH genotypes had 1.8 times the shoot biomass in the minimum tillage treatment, where soil strength is greater, compared with that in the conventional tillage treatment. In contrast, the genotypes with root hairs had the same biomass production in both tillage treatments.

Fig. 2.

(a) Total shoot dry mass (g), (b) shoot P concentration (mg g^–1), and (c) total shoot P (mg) of root hairless (NRH), root hair (RH), and wild-type (WT) barley genotypes grown for 8 weeks in soil cores taken from a conventional tillage and minimum tillage treatment of an 8-year-old field-based tillage trial. Bars show the least significant difference (P < 0.05; n=15) for the interaction between tillage treatment and root hair genotype.

P uptake

There was a significant three-way interaction between soil bulk density, P availability, and genotype on shoot P concentration, total shoot P, and P uptake per unit root length (i.e. the parameters of P uptake; Table 3 and Fig. 1). In P0 soil, the values of the parameters for P uptake of the RH genotype were >3-fold larger than those of the NRH genotype, regardless of soil density (P < 0.05). Furthermore, there was no effect (P > 0.05) of increasing soil bulk density on the value of any of these parameters of P uptake, except for a marginal increase in P uptake per unit root length (P < 0.05) between 1.2g cm^–3 and 1.6g cm^–3 for the RH genotype. In the P0 soil, the average shoot P concentrations were within the deficient to marginal range (1.21mg g^–1 and 3.75mg g^–1) for the RH and NRH genotypes, respectively (Reuter and Robinson, 1986).

In the P500 soil, shoot P concentration, total shoot P, and P uptake per unit root length were at least 3-fold higher than those of plants grown in P0 soil (main effect of P; Table 3 and Fig. 1). Average shoot P concentrations ranged from 5.07mg g^–1 to 8.61mg g^–1, which is within the adequate range (Reuter and Robinson, 1986). In contrast to the P0 soil, the value of all of the parameters of P uptake increased significantly with increasing soil bulk density (i.e. P uptake increased between 1.2 and 1.6 and/or 1.7g cm^–3). Furthermore, in the P500 soil, the P uptake of the NRH genotype increased relative to that of the RH genotype as the bulk density of the soil increased. At a bulk density of 1.2g cm^–3, the RH genotype had double the shoot P of the NRH genotype. However, at 1.6g cm^–3, shoot P of both genotypes had increased (by 25% and 150% for the RH and NRH genotypes, respectively) and did not differ significantly between the two genotypes (Fig. 1a). A similar trend was observed for shoot P concentration and P uptake per unit root length; however, for these two parameters, there was a relatively small difference (P < 0.05) between the genotypes at 1.7g cm^–3. These trends were also seen when plants were grown in cores taken from a field-based tillage trial (Fig. 2). Again the NRH accumulated less P than the RH genotypes when the soil bulk density was less in the conventional tillage, but accumulated the same amount of P as the RH genotype when the soil bulk density was higher in the minimum tillage treatment.

Root parameters

Root hairs and rhizosheaths were observed on both the NRH and RH barley lines. Root hairs on the NRH genotype were significantly shorter than those of the RH genotype, with average lengths of 0.06mm and 0.71mm, respectively (Table 2). This was also reflected in smaller rhizosheaths (mass of dry soil per unit root length) on the NRH genotype. Increasing soil density resulted in decreased root hair length of the RH genotype by an average of 20% (1.6g cm^–3 or 1.7g cm^–3 relative to 1.2g cm^–3). Root hairs of the RH genotype also responded to application of P, with root hairs 19% shorter in the P500 treatment relative to the P0 treatment. Similar responses were observed for the rhizosheath, with smaller rhizosheaths observed at higher soil bulk density and at high P supply for the RH genotype. While there was no effect of soil density or P application on the root hair length of the NRH genotype, rhizosheaths on the NRH genotype were generally smaller in the low-strength soil.

The average root diameter across the whole root system of the NRH genotype was ~20% smaller than that of the RH genotype (P < 0.001; Table 2). In the P0 soil, the NRH genotype maintained smaller average root diameters than that in the P500 soil (1.2g cm^–3 and 1.6g cm^–3) even as root diameter increased with increasing soil bulk density (three-way interaction P < 0.10).

Discussion

The results indicated important significant interactions of root hairs with bulk density and P supply, including effects on soil penetration. The results will be discussed in the order of the experiments.

Root hairs and root penetration

Root hairs offered advantage for root penetration into high-strength layers. Given that the two genotypes did not differ in the number of seminal roots per plant (and hence anchorage provided by these seminal roots), root hairs considerably facilitated root penetration into high-strength layers. A number of chemical, physical, and biological factors in the rhizosphere are likely to be associated with root hairs aiding the penetration of roots into high-strength layers. Theoretical calculations indicate that root hairs can provide adequate tensile strength to anchor root tips (Bengough et al., 2011). Several studies have also demonstrated that mucilage and/or microbes in the rhizosphere adhere soil to roots (Watt et al., 1993; Czarnes et al., 1999). This is likely to be associated with the development of the rhizosheath, which itself is linked with the development of root hairs, amongst other processes. This has obvious advantages in agricultural systems with high-strength layers at shallow depth (plough pans, traffic compaction, and droughted topsoil), allowing plants the ability to penetrate through and acquire resources from deeper layers.

The main advantage of root hairs for helping roots to penetrate into high-strength layers appears to be in increasing the number of roots that can penetrate into high-strength layers, rather than improving the elongation rate once in the high-strength layer. Furthermore, when the two genotypes were grown in soils of uniform high strength, differences in root length were not detected between the NRH and RH genotypes (Experiment 2). Root hairs probably confer most advantage for root penetration and growth from a low-strength region of soil into a high-strength region given that root hair length would be restricted in a uniformly high-strength soil. The degree of advantage remains to be determined in field soils with a range of soil structures, or within soils where root growth is mainly confined to biopores.

Root hairs, P uptake, and soil strength

The value of root hairs for P uptake under combined soil stresses was evident in the present experiments. The results confirm the value of root hairs for P uptake in P-deficient soils and further demonstrate their value for P uptake in high-strength soils. Despite an ~20% reduction in root hair length in a high-strength soil, the genotype with root hairs maintained a significant advantage for P uptake over the NRH genotype in high-strength soil. In P-sufficient conditions, the value of root hairs for P uptake was reduced in high-strength soil. The results suggest that under certain soil physical conditions, root hairs are a highly effective strategy for P acquisition.

The present results showed improved P uptake (per unit root length) in high-strength soils. Earlier research by Cornish et al. (1984) attributes such a result to improved root–soil contact rather than increased diffusion of P to the root. This improved P uptake could be due to factors including: a larger root diameter in high-strength soil effectively increases the surface area of the root per unit root length (though this response reduces specific root length and hence makes it a less efficient allocation of carbon); the higher density of soil increases the quantity of P within close proximity of the root; and increased unsaturated hydraulic conductivity allows greater movement of P. Hence, the decreased cylinder of soil that the roots could explore (shorter root hairs and/or shorter roots) was compensated by the increased mass of soil and P within close proximity to the root. As Cornish et al. (1984) notes, the absolute effect of soil strength on total P uptake will depend on the relative magnitude of these two opposing effects.

It is interesting to consider the effect of root diameter on root–soil contact further. The NRH genotype had roots that were thinner than those of the RH genotype. Increasing specific root length (i.e. increased root length per unit root mass) is a common response in P-deficient soils and may allow plants to allocate carbon for P uptake more efficiently (Eissenstat, 1992; Horst et al., 1993; Gahoonia and Nielsen, 2004b). However, this means that for a given unit of root length, the NRH genotype had less surface area. Furthermore, due to P deficiency or high soil strength, the NRH plants did not compensate for this reduced surface area per unit root length by increasing root length. Given that P uptake per unit root length of the NRH genotype was equivalent to that of the genotype with root hairs in the high available P soil (1.6g cm^–3), the data suggest that the benefit to P uptake in high-strength soil is largely due to the increased mass of soil and hence more P within close proximity to the root. This improved root–soil contact might explain the variation seen in barley in its ability to acquire P in minimum tillage compared with conventional tillage systems (George et al., 2011). The overall better shoot growth and P uptake of plants in the high-strength soils might also relate to the overall larger mass of soil in the pots and hence overall greater supply of water and nutrients.

An unexpected finding in this work was the presence of root hairs on the NRH genotype. The root hairs were very short (<0.1mm) and may reflect elongation mutants (Schiefelbein and Somerville, 1990). It was not possible to establish whether the initiation of root hairs was also affected, as the very short length of root hairs in NRH genotypes precluded this. These short root hairs might not have been detected (or perhaps were not even present) in the initial screen of the mutant population, which was conducted by photographing germinated seeds on filter paper in the absence of any mechanical impedance from soil (Gregory et al., 2009). However, this appears unlikely given that root hairs have not previously been observed on this specific NRH line when studied under a microscope (Brown et al., 2012). Furthermore, root hairs on the NRH genotype were distinct from the previously reported ‘bulges’ (<0.01mm length) that were found covering the surface of some genotypes by Brown et al. (2012). The occurrence of short root hairs on genotypes that previously did not exhibit this genotype could be related to the fact that the genotypes were not isolines and the mutational effect is becoming less pronounced with regeneration of the lines, or that mechanical impedance to the root tip may promote root hair elongation as observed previously in barley (Goss and Russell, 1980). Regardless, the contrasting root hair lengths still provide useful treatments for quantifying the impact of root hairs on P acquisition.

Value of root hair traits in the field

Few studies have examined the value of specific root traits to improve P uptake under multiple soil stresses. In many cases, field trials may have inadvertently evaluated traits or genotypes for performance in high-strength soils, given the prevalence of this constraint after seedbed ageing (Hallett and Bengough, 2013). In this work, the value of root hairs for P uptake in high-strength soils was specifically demonstrated. The results also give an indication that supply of P to the root can be manipulated by several different mechanisms. Supply of P to the root can be manipulated by directly increasing supply through application of fertilizer, increasing root contact with the soil by increasing the surface area of the root (i.e. root hairs), and/ or increasing root contact with the soil by increasing the bulk density of the soil. The ability of a genotype to take advantage of these various mechanisms may explain the genotypic variation seen in barley populations to acquire P in a range of tillage treatments (George et al., 2011), where few genotypes were capable of acquiring P efficiently in systems with both strong and loose soils.

A significant driver for the adoption of minimum tillage and controlled-traffic operations is reduced compaction by machinery. While this has had the anticipated benefit of improving overall soil structure, many studies report increased soil bulk density or greater soil strength under minimum (or no) tillage operations relative to cultivated soils (Lampurlanes and Cantero-Martinez, 2003; Osunbitan et al., 2005; Morell et al., 2011). In the intact core experiment (Experiment 3), the improved performance of the NRH genotype in the minimum tillage treatments probably reflects improved root–soil contact under higher strength conditions. In contrast, genotypes with root hairs maintained shoot dry matter and shoot P under both conventional and minimum tillage conditions. It is possible that root hairs are an adaptation to low-strength soils (i.e. loose or cultivated soils) and hence a trait that has been favoured under the selection conditions of modern plant breeding. However, it is also likely that root hairs are advantageous in uncultivated soil where roots are confined to biopore channels.

The ability of root hairs to improve root penetration into high-strength soil layers further demonstrates that root hairs are a valuable trait for crop growth. In the field, the ability to improve root penetration is important for increasing the overall size of the root system and hence access to soil resources, particularly resources at depth, but it may also be crucial for plant establishment in a dry seedbed. For example, access to subsoil water has been shown to be particularly important in water-limited environments as relatively small amounts of subsoil water can have large impacts on grain yield (Kirkegaard et al., 2007). Furthermore, the ability of root hairs to improve root anchorage might also have advantages for roots growing in the soil macrostructure. Many roots, particularly at depth, are found growing in the soil macrostructure (White and Kirkegaard, 2010). Root hairs might help to anchor these roots in the pore space and hence aid root growth through pores, root adherence to the pore wall, and hence transport between the pore wall and root surface. This research highlights the need to consider multiple stresses when assessing the usefulness of root traits for improved agricultural sustainability. Consideration of combined stress is critical when an attempt is made to design plants with appropriate root systems for a second green revolution.

Conclusions

High soil strength or compaction is a widespread constraint that restricts root and rhizosphere development. In this work, it was demonstrated that despite a reduction in root hair elongation under high-strength conditions, root hairs remain a valuable trait for P uptake in P-deficient, high-strength soils. Under high P conditions, improved root–soil contact coupled with an increased supply of P to the root may be responsible for reducing the value of root hairs for P uptake in high-strength soils relative to loose soils. Furthermore, root hairs may also help to anchor roots in the soil and aid root penetration into high-strength layers. In conclusion, root hairs confer an advantage to crop growth and resource-use efficiency under multiple soil constraints. Further work is required to investigate whether genotypes selected for longer root hairs under low soil strength conditions maintain a relative advantage for root hair length and P uptake under high soil strength.

Acknowledgements

The authors would like to thank Bruna Arruda and Joice Heidemann for assistance with the experimental work. This work was funded by the Scottish Government through work package 3.3 and the University of New England Early Career Post-doctoral initiative.