Rooting patterns of four crop legumes in response to seed-placement depths in the dry season

Abstract

On sandy paddy fields, key factors for successful crops in the dry season without irrigation are a shallow water table and practices such as deep seed-placement but only some legume species are adapted to such conditions. To understand the adaptation of legume species to deep seed-placement over shallow water tables, we studied their rooting patterns on two sandy soils. Cowpea (Vigna unguiculata), mungbean (Vigna radiata), peanut (Arachis hypogaea) and soybean (Glycine max) seeds were sown shallow (∼5 cm) or deep (∼15 cm) in deep sandy soils after harvesting rice in two shallow water table locations in north-east Thailand. The legumes depended mainly on capillary water rising from the water table and none experienced water deficit throughout the growing season. Generally, deeper seed-placement decreased overall root dry weight, but it increased the root surface area to weight ratio. Deep seed-placement promoted a greater fraction of root growth into the subsoil for cowpea (86–99% of total root length), mungbean (61–93% of total root length) and peanut (78–98% of total root length) where the soil contained more water throughout the growing season. Moreover, deep seed-placement at the site with the lower water table promoted deeper penetration of roots of cowpea (∼20 cm deeper), mungbean (∼20–40 cm deeper) and peanut (∼20–40 cm deeper) which improved water access, especially late during the growing season when topsoils dried to close to wilting point. Unlike other species, the soybean rooting pattern did not respond much to seed-placement depths, or soil moisture.

Keywords:

• Cowpea
• mungbean
• peanut
• rooting system
• seeding depth
• soybean

Introduction

There are numerous rainfed cropping systems based on planting dry-season crops without any irrigation after harvesting rice (e.g. Thailand – Polthanee 1991, Nigeria – Adigbo et al. 2007, Indonesia and the Philippines – Rahmianna et al. 2000, So and Ringrose-Voase 2000). To support growth and significant crop yields, such dry-season crops are planted on specific locations with access to water sourced from residual soil moisture and additional water from capillary rise plus occasional out-of-season rainfall. In sandy paddy fields, key factors which could help crops grow successfully in the dry season without irrigation are a shallow water table and some agricultural practices such as deep seed-placement but nevertheless only some legume species are adapted to such conditions (Polthanee 1991).

In the dry season when the topsoil dries out, increased seed-placement depth may improve crop establishment when additional moisture is available in the subsoil (Siddique and Loss 1999). It may lead to root penetration to greater soil depth to explore a large volume of soil water that remains available throughout the crop cycle (Polthanee 2001). Ehsanullah et al. (1999) found that wheat (Triticum aestivum) sown at 5 cm gave higher root growth rate and root–shoot ratio than wheat sown at 3 cm under rainfed conditions when soil moisture was available at depth. Root growth and root–shoot ratio are important for rainfed wheat because higher values of both induced drought tolerance in crop plants (Ehsanullah et al. 1999). However, excessively deep seed-placement (>10 cm) could also reduce growth and yield of crops such as peanut (Rao and Reddy 1985, Nambiar and Srinivasa Rao 1987) even though it places the seed closer to the stored moisture in the soil profile. In addition, seedlings from deeper seed-placement showed greater dry weight allocation to the stem, at the expense of the roots (Seiwa et al. 2002, Gholami et al. 2007).

Root mass and rooting depth of crops planted in the dry season without irrigation play an important role for crop survival because the water table in the dry season increases in depth with time (Passioura 1983, Turner 1986, Ludlow and Muchow 1990, Meisner and Karnok 1992, Ketring and Reid 1993, Sangakkara et al. 2001, Matsui and Singh 2003, Taiz and Zeiger 2006, Kashiwagi et al. 2008). Deep seed-placement gives the roots earlier access to moisture in the deeper soil (Polthanee 1983). Chantron (1983) suggested that deep seed-placement allows the root system to utilize the soil water at greater depth more efficiently. However, there is still limited direct evidence to support this suggestion.

Among legume crop species, different patterns of root distribution including greater root length and/or density in deeper soil layers might be associated with their survival in the dry season without irrigation. Pandey et al. (1984b) reported that cowpea and peanut had higher root densities at 0.4 to 0.8 m soil depths than soybean or mungbean. Peanut and cowpea roots also had greater ability to extract water from deeper soil to maintain an adequate water uptake. In contrast, soybean and mungbean had relatively shallow root systems and did not maintain water uptake under the drier soil moisture regime (Pandey et al., 1984b). In addition, Benjamin and Nielsen (2006) reported that field pea (Pisum sativum L.) and chickpea (Cicer arietinum L.) had a greater proportion of their root systems deeper in the soil profile than soybean, which could lead to better use of stored soil water.

The present experiment was set up to investigate the differences of rooting patterns among four legume crop species sown at different depths of seed placement in the dry season on two sandy soils without any irrigation water supply.

Materials and methods

Experimental design and treatments

The experiment was conducted under the farmers’ field conditions at Kokeyai village (Baan Fang district) (latitude 16° 28′ N, longitude 102° 39′ E) and Samjan village (Muong district) (latitude 16° 42′ N, longitude 102° 48′ E) located in Khon Kaen Province, north-east Thailand during 13 December 2007 to 27 March 2008 (dry-season). The soil in Kokeyai was a Typic Paleustult and in Samjan was an Arenic Haplustalf. Soils were analysed for soil texture by the hydrometer method, for pH by 1:2.5 H₂O, organic matter by Walkley and Black wet oxidation, total N by a micro-Kjeldahl method, extractable P by Bray II and exchangeable K and Ca by using 1 M ammonium acetate extraction at pH 7. Physical and chemical properties of soil in Kokeyai and Samjan villages are presented in Table I. Both locations were selected because they experience capillary water rise from a shallow water table in the dry season. Rate of lateral flow of shallow water table in the fields of Khon Kaen province varied between 13.6–15.8 mm day⁻¹ for the outflow and between 4.7–9.2 mm day⁻¹ for the inflow (Tsubo et al. 2007) and hence were not likely to affect water supply to the experimental site. Saturated hydraulic conductivity (K_sat) of this soil type in north-east Thailand can vary from 29 to 103 cm day⁻¹ (Trelo-ges and Sriboonlue 2002). Legume seeds were sown at two depths, approximately 5 cm (called shallow seed-placement) which was the common seed-placement (Dungan and Ross, 1957, Martin and Leonard 1965) and approximately 15 cm (referred to as deep seed-placement). We selected cowpea (Vigna unguiculata c.v. ‘KKU 264 R’) and peanut (Arachis hypogaea c.v. ‘Tainan 9’) as test species which previously grew well with deep seed-placement and no irrigation in farmers’ fields, plus mungbean (Vigna radiata c.v. ‘U-thong 1’) and soybean (Glycine max c.v. ‘NakhonSawan 1’) which did not grow well in the dry season without irrigation (Pandey et al. 1984b).

Table I. Soil physical and chemical properties of profiles (0–100 cm) at experimental sites in Kokeyai and Samjan village, KhonKaen province, Thailand in dry season, 2007–2008.

	Sand	Silt	Clay			OM	Total N	Extr P	Extr K	Extr Ca
Soil depth (cm)	(%)			Texture class	pH	(%)			(mg kg^−1)
Kokeyai
0–15	49	40	11	Loam	6.09	3.19	0.16	13	19	394
15–30	43	41	16	Loam	6.35	2.81	0.14	6	13	612
30–45	47	41	12	Loam	6.42	0.61	0.03	3	12	260
45–60	44	43	14	Loam	5.29	1.68	0.08	5	10	445
60–100	50	38	12	Loam	6.40	0.66	0.03	3	11	278
Samjan
0–15	75	9	16	Sandy loam	6.39	1.67	0.08	5	37	941
15–30	82	9	9	Loamy sand	6.59	1.01	0.05	4	15	470
30–45	82	10	7	Loamy sand	6.56	0.63	0.03	3	12	277
45–60	82	12	6	Loamy sand	6.50	0.28	0.01	3	9	298
60–100	83	10	7	Loamy sand	6.38	0.66	0.03	5	11	251
OM, Organic matter.

Crop management

After rice was harvested from each field, rice straw was removed and following farmer practice, soil was ploughed and harrowed three times to loosen the top soil to 15 cm depth. The soil was allowed to dry for 2 to 3 days after each round of ploughing and harrowing. Before final ploughing and harrowing, 15-15-15 (N: P₂O₅: K₂O) fertilizer was applied to soil at 156 kg ha⁻¹ and lime applied at 625 kg ha⁻¹ to raise pH to nearly 7.

For shallow seed-placement, seeds were placed in furrows made by a Planet Jr. seed drill (approximately 5 cm from the soil surface). For deep seed-placement, manual seed-placement of seeds was accomplished using a plough drawn by a two-wheel tractor to make furrows approximately 15 cm depth. The seeds were then dropped in the furrows and covered with the soil when the adjacent furrow was prepared. Seeds were soaked in water for 24 hours before planting and treated with Captan (3a,4,7,7a-tetra-hydro-2-[(trichloromethyl) thio]-1H-isoindole-1, 3(2H)-dione) at 5 g kg⁻¹ seed to protect seedlings from crown rot. Seeds were sown with 2 seeds per hill in 0.5 metre row-to-row spacing and 0.10 metre plant-to-plant spacing. There was no weeding, or fertilizer application after planting. Pests and diseases were adequately controlled during the growing season. The experiment was kept under rainfed conditions, so there was no irrigation in the field. There were no adjacent crops in both experimental fields during the growing season.

Data collection

Soil moisture and water table

Soil moisture was recorded by using the gravimetric method at the depths of 0–15, 15–30, 30–45, 45–60 and 60–100 cm. Soil samples were collected one core per plot from each soil layer by soil auger at 0, 2, 5, 7, 12 and 14 weeks after seed-placement (WAS). The soil samples were weighed before and after oven drying at 105 °C for 48 hours. Water table level was observed by the rise of water in perforated PVC tubes, installed at 20-, 40- and 100-cm soil depth. Water table depth was calculated from the soil surface to water level distance by dropping a calibrated wooden stick into the PVC tube throughout the experiment.

Root sampling

Legume roots were sampled at 3, 6 and 9 WAS from each plot. Roots were sampled using a coring tube (Welbank et al. 1974) of 76 mm diameter and 1.15 m length. Legume shoots were excised at the soil surface before root sampling. In each plot, one sampling core was centred over the tap root and the second in soil between the plants in the row. Roots were extracted from the soil by driving the sampling core into the soil by a jackhammer and removing it using a lever. The soil core was sectioned at 0–5, 5–15, 15–30, 30–45, 45–60, 60–80 and 80–100 cm from the soil surface. The samples were placed in sealable plastic bags and stored under refrigeration. The following day, roots were washed from the soil cores on 0.2 mm sieves and measured for root length and root surface area using a scanner to record the root data which were then calculated by the WINRHIZO Pro 2004a software (REGENT Instruments Inc., QC, Canada). Root length density (RLD) was calculated as:

Average root length densities were calculated from means of root length densities over the tap root and in the row between the plants.

After root scanning, roots were dried in an oven at 80 °C for 48 hours and the dry weights were measured.

Seed yield

After physiological maturity, seed yield were determined from 50 hills randomly chosen in each plot. Pods were shelled out and cleaned after air drying. Mature, intact and healthy seeds were then weighed for yield calculation.

Statistical analysis

A split-plot design was used to examine the effect of seed-placement depth on root dry weight, root length, root surface area and seed yield of the four species. To test for effects of root depth and species on root length density, repeated measured design was used with species and seed-placement depth as main factors and root depth as a repeated measure. Each sampling date was analysed as a separate experiment. Calculation procedures were done using STATISTIX-8.

Results

Rainfall and water table depth

During the growing season, rainfall at both locations was minimal. The crops received rainfall in Kokeyai twice at 7 (8.5 mm) and 13 (24.3 mm) WAS and Samjan received rainfall at 6 (4.8 mm), 7 (2.7 mm) and 13 (7.8 mm) WAS. Kerdsuk (1986) found that peanut after rice in the dry season without irrigation required 474 mm of water for growth, while stored soil water to 1 m depth at planting plus rainfall during the growing season was 295 mm and there was 151 mm of residual soil water at harvesting. Apart from the residual soil moisture left over from the preceding rainy season, capillary water rise from a shallow water table was the main factor controlling soil water supply and hence growth and yield of peanut after rice (Kerdsuk 1986). Thus, crop growth in this study probably depended mainly on capillary water rising from the water table, apart from the residual soil moisture present in the soil profile at sowing.

The measured water table depths dropped from 69 to 133 cm below the soil surface in Kokeyai village and from 28 to 102 cm below the soil surface in Samjan village (Figure 1). The water table depth declined below 100 cm at about 9 WAS in Kokeyai and 12 WAS in Samjan. Thus, crops in Samjan had access to more water than crops in Kokeyai.

Figure 1.  Weekly rainfall and water table depth at experimental sites in Kokeyai (a) and Samjan (b) villages.

Soil water content

Soil water contents in Kokeyai were lower than those in Samjan at 0–15, 15–30, 30–45, 45–60 and 60–100 cm soil depths (Figure 2) presumably because of the lower water table depth in Kokeyai. At every soil depth, soil water content of both locations remained in the plant-available ranges throughout the growing season. Soil water content in the topsoil of deep seed-placement was lower than shallow seed-placement and this was attributed to the break in capillary rise of water caused by deep ploughing. Late in the growing season, soil water content in the topsoil (0–15 cm from soil surface) at Kokeyai decreased almost to the permanent wilting point (PWP). In Kokeyai, shallow-placed plots had higher soil water content than deep-placed plots at every soil layer. At Samjan, soil water content of shallow- and deep-placed plots were not different possibly because this location had a shallow water table and abundant soil water.

Figure 2.  Soil moisture content of shallow (▪) and deep seed-placement (□) treatments at 0–15 cm (a), 15–30 cm (b), 30–45 cm (c), 45–60 cm (d) and 60–100 cm (e) in Kokeyai and Samjan experimental sites. Horizontal lines represent water contents at field capacity (FC) and permanent wilting point (PWP). Values are means of four replications.

Average root dry weights

Average root dry weights (RDW) were highest at both locations in cowpea and at every sampling date except in Kokeyai at 3 WAS (Table II). In general, peanut had lower RDW than cowpea but higher RDW than mungbean and soybean. Legumes after shallow seed-placement had 42–145% higher RDW than deep seed-placement in both locations and all sampling dates (Table II).

Table II. Average root dry weight, root length density, root surface area in 1 m soil profile and seed yield of cowpea, mungbean, peanut and soybean with shallow and deep seed-placement in Kokeyai and Samjan experimental sites at 3, 6 and 9 weeks after sowing (WAS).

		Average root dry weight (g m^−3)						Average root length density (×10^−2 cm cm^−3)						Average root surface area (×10^−2 cm^2 cm^−3)						Yield (kg ha^−1)
		Kokeyai (WAS)			Samjan (WAS)			Kokeyai (WAS)			Samjan (WAS)			Kokeyai (WAS)			Samjan (WAS)
Legume	Seed-placement depth	3	6	9	3	6	9	3	6	9	3	6	9	3	6	9	3	6	9	Kokeyai	Samjan
Cowpea	Shallow	15.6	122.1	221.2	11.7	45.1	88.8	6.37	16.77	20.25	4.09	5.82	10.15	0.88	3.08	3.24	0.61	1.31	1.71	536	400
Cowpea	Deep	13.2	48.1	72.3	10.1	25.5	32.6	6.72	12.63	20.52	4.84	7.11	11.34	0.99	1.83	2.82	0.70	0.80	1.63	185	452
Mungbean	Shallow	13.2	36.8	45.5	2.5	20.7	68.6	3.97	9.89	12.72	1.98	8.76	6.19	0.49	1.29	1.51	0.26	0.92	1.04	933	1069
Mungbean	Deep	5.8	19.9	17.8	3.1	14.0	18.8	3.61	9.91	9.21	2.95	7.21	5.94	0.41	1.11	1.13	0.36	0.78	0.96	972	915
Peanut	Shallow	23.3	45.2	101.9	13.0	35.4	61.2	9.45	12.11	14.81	4.38	8.01	8.53	1.30	1.86	2.19	0.62	1.13	1.30	777	469
Peanut	Deep	12.5	38.7	66.7	6.1	21.5	25.0	6.11	16.96	18.86	2.66	4.85	7.88	0.84	1.91	2.36	0.40	0.61	1.04	840	433
Soybean	Shallow	14.5	28.6	42.7	10.2	35.3	35.7	9.90	8.60	12.40	4.88	8.45	6.82	1.13	1.11	1.39	0.57	0.99	0.87	701	459
Soybean	Deep	13.0	28.3	39.7	7.1	16.7	27.8	8.04	12.87	7.80	3.86	6.13	7.13	0.94	1.48	0.85	0.51	0.77	0.80	896	428
SED
Legume (L)		3.4^ns	6.2***	8.4***	1.3**	2.4***	7.5*	1.27^ns	1.05**	1.99**	0.70^ns	0.89^ns	2.00^ns	0.17*	0.16***	0.24***	0.09**	0.12^ns	0.32^ns	124**	78***
Sowing depth (D)		1.1***	5.1***	4.5***	0.6**	2.1***	3.4***	0.64^ns	0.88^ns	1.31^ns	0.27^ns	0.57*	0.72^ns	0.16^ns	0.14^ns	0.19^ns	0.04^ns	0.07***	0.10^ns	134^ns	85^ns
L×D		3.8*	10.0**	10.5***	1.6**	3.8^ns	8.8**	1.56^ns	1.63*	2.72^ns	0.79*	1.20^ns	2.25^ns	0.20^ns	0.25**	0.36^ns	0.11^ns	0.15^ns	0.35^ns	190^ns	120^ns
SED, Standard error of the difference between means (p < 0.05).
ns, Non- significant, *, **, *** Significant at the 0.05, 0.01 and 0.001 probability levels, respectively.

There were interactions between legume species and seed-placement depths for RDW in every location and sampling date except in Samjan at 6 WAS (Table II). At 3 WAS, peanut after shallow seed-placement had highest RDW in both locations, but mungbean after deep and shallow seed-placement had lowest RDW in Kokeyai and Samjan, respectively. However, RDW of cowpea after shallow seed-placement were highest and mungbean after deep seed-placement were lowest at 6 WAS in Kokeyai and 9 WAS in both locations.

Average root length density

Average root length density (RL) of legume species was significantly different in Kokeyai at every sampling date and in Samjan at 3 WAS (Table II). In Kokeyai, RL of soybean, peanut and cowpea were higher than mungbean at 3 WAS but at 6 WAS, cowpea and peanut had greater RL than soybean and mungbean. At 9 WAS, RL of cowpea was higher than peanut, mungbean and soybean, respectively. Meanwhile in Samjan at 3 WAS, cowpea and soybean had higher RL than peanut and mungbean, respectively (Table II). At 6 WAS, cowpea RL like RDW was greater than peanut which in turn was greater than mungbean and soybean.

Legumes after shallow seed-placement had higher RL than deep seed-placement in Kokeyai at 3 WAS and in Samjan at 6 WAS. In Kokeyai after 6 WAS, deep-placed peanut had greatest RL but soybean after shallow seed-placement had lowest RL. In Samjan after 3 WAS, RL of shallow-placed soybean was highest but shallow-placed mungbean was lowest (Table II).

Average root surface areas

In Kokeyai, peanut, soybean and cowpea had greater average root surface area (RSA) than mungbean at 3 WAS but at 6 and 9 WAS, RSA of cowpea were higher than peanut, mungbean and soybean, respectively. In Samjan at 3 WAS, cowpea had highest RSA while mungbean had the smallest RSA.

Shallow-placed legumes had more RSA than deep-placed legumes in Kokeyai at 3 and 6 WAS and in Samjan at 6 WAS. There was no interaction between legume species and seed-placement depths for RSA except in Kokeyai at 6 WAS and in Samjan at 3 WAS (Table II).

Seed yield

Cowpea had lowest seed yield in Kokeyai and mungbean had highest yield in Samjan (Table II). Seed yields of legume after shallow and deep seed-placement were not significantly different (Table II). There was no interaction between legume species and seed-placement depth (Table II).

Root length density distributed in 1 m soil profile

Root length densities (RLD) of legume species distributed in 1 m soil profile were significantly different in Kokeyai at 6 and 9 WAS and in Samjan at 3 WAS (p < 0.09) (Table III). For seed-placement depths, RLD were significantly different in both locations and at all sampling dates except in Kokeyai at 6 WAS (Table III). Meanwhile, root depth caused significant differences in RLD distribution in every location and sampling date. The interactions between seed-placement depth and soil depth for RLD were significant in both locations at every sampling date (Table III).

Table III. p-value of root length density (RLD) of cowpea, mungbean, peanut and soybean with shallow and deep seed-placement at 3, 6, and 9 weeks after sowing (WAS) in Kokeyai and Samjan.

		Kokeyai (WAS)			Samjan (WAS)
SOV	DF	3	6	9	3	6	9
Legume (L)	3	0.329	0.054	0.002	0.085	0.650	0.490
Error (a)	9
Seed-placement depth (D)	1	0.007	0.114	0.020	0.028	0.029	0.055
L×D	3	0.969	0.349	0.495	0.304	0.672	0.834
Error (b)	12
Root depth (R)	6	0.000	0.000	0.000	0.000	0.000	0.000
L×R	18	0.628	0.061	0.115	0.011	1.000	0.133
D×R	6	0.000	0.000	0.000	0.000	0.000	0.000
L×D×R	18	0.990	0.000	0.467	0.084	0.334	0.803
Error (c)	144
Total	223
SOV, source of variation; DF, degree of freedom.

Legumes in Kokeyai had larger root systems than in Samjan especially late in the growing season. With the exception of soybean, their roots penetrated deeper in Kokeyai where the soil was drier than at Samjan (Figures 3 and 4). Root mass of soybean in the late growing season (9 WAS) was greater than in the middle growing season (6 WAS). However, its root depths in the late and in the middle growing season were similar. In general, very few soybean roots grew below 45 cm especially with deep seed-placement (Figures 3 and 4).

Figure 3.  Root length density (RLD) by depth for cowpea, mungbean, peanut and soybean with shallow and deep seed-placement at 3, 6, and 9 weeks after sowing (WAS) in Kokeyai. Statistical tests for the data are shown in Table III. Dashed line represents the water table depth.

*, Root distribution (RD) at 0–15, 15–45, 45–60 and 60–100 cm soil layer (percentage of total length from 0–100 cm soil layer).

Figure 4.  Root length density (RLD) by depth for cowpea, mungbean, peanut and soybean with shallow and deep seed-placement at 3, 6 and 9 weeks after sowing (WAS) in Samjan. Statistical tests for the data are shown in Table III. Dashed line represents the water table depth.

*, Root distribution (RD) at 0–15, 15–45, 45–60 and 60–100 cm soil layer (percentage of total length from 0–100 cm soil layer).

In Kokeyai at 3 WAS, roots of legume after shallow and deep seed-placement penetrated to the 45–60 cm soil layer (30–45 cm from seed position) except cowpea and mungbean after shallow seed-placement that penetrated to only 30–45 cm (25–40 cm from seed position) (water table depth at 3 WAS was approximately 86 cm below the soil surface) (Figure 3). At 6 WAS, after deep seed-placement roots penetrated to 60–80 cm soil layer (45–65 cm from seed position). After shallow seed-placement, mungbean and soybean roots penetrated to 45–60 cm (40–55 cm from seed position) and cowpea and peanut roots penetrated to 60–80 cm (55–75 cm from seed position) (water table depth at 6 WAS was about 94 cm below the soil surface). When water table depth was approximately 100 cm at 9 WAS, roots of legumes after shallow and deep seed-placement penetrated to 60–80 (55–75 cm from seed position) and 80–100 cm soil layer (65–85 cm from seed position), respectively, except soybean that reached only to 45–60 cm (40–55 cm from seed position) after shallow seed-placement and 30–45 cm (15–30 cm from seed position) after deep seed-placement (Figure 3).

At 3 WAS, when the water table was about 51 cm below the soil surface in Samjan, legume roots after shallow and deep seed-placement penetrated to 15–60 cm soil layer (10–55 cm from seed position) and 30–80 cm (15–65 cm from seed position), respectively (Figure 4). At 6 WAS, roots of shallow- and deep-placed legume penetrated to the 30–80 cm soil layer (25–75 cm from seed position) and 45–80 cm soil layer (30–65 cm from seed position), respectively, when the water table was about 58 cm from the soil surface. Legume roots penetrated to the 45–80 cm soil layer after shallow (40–75 cm from seed position) and deep seed-placement (30–65 cm from seed position) at 9 WAS when water table depth was about 83 cm (Figure 4).

Most roots of cowpea, mungbean, peanut and soybean established by shallow seed-placement were distributed in the topsoil (0–15 cm from soil surface) but with deep-placed seeds most roots were distributed in subsoil (15–100 cm from soil surface) in both locations (Figures 3 and 4). Peanut after shallow seed-placement had about 70% of the RLD in the topsoil in Kokeyai and 86% in Samjan. By contrast, with deep seed-placement, about 2% of the RLD was in the topsoil in Kokeyai and 15% in Samjan. Root LD percentages of cowpea with shallow seed-placement in the topsoil were 57 to 91% and 84 to 96% in Kokeyai and Samjan, respectively. However, after deep seed-placement only 1 to 14% and 5 to 14% of RLD was in topsoil in Kokeyai and Samjan, respectively. For mungbean, percentages of RLD in topsoil ranged between 56 to 84% and 59 to 93% in Kokeyai and Samjan, respectively, with shallow seed-placement. Percentages of RLD in the subsoil ranged between 16 to 44% and 7 to 41% for the shallow-placed mungbean and 85 to 93% and 61 to 93% for the deep-placed mungbean in Kokeyai and Samjan, respectively (Figures 3 and 4). Root LD of shallow-placed soybean in topsoil ranged between 82 to 93% in Kokeyai and 70 to 89% in Samjan and deep-placed soybean ranged between 19 to 24% and 8 to 26% in Kokeyai and Samjan, respectively (Figures 3 and 4).

Legume roots were mostly distributed in the soil layer adjacent to the seed position, which were the 5–15 cm soil layer for shallow seed-placement and 15–30 cm soil layer for deep seed-placement. Roots of legumes after shallow and deep seed-placement that grew above the seed position during the growing season comprised between 0.3–1.9% and 8.8–13.2% of the root system in Kokeyai and 0–1.2% and 10.7–26.2% in Samjan, respectively (Figures 3 and 4).

Root surface area to weight ratio

All legumes after deep seed-placement had higher root surface area to weight ratio (AWR) than shallow-placed legumes regardless of time during the growing season or location (Figure 5). Legumes at 3 WAS had higher AWR than that at 6 and 9 WAS.

Figure 5.  Root surface area to root weight ratio in the 0–1 m depth at 3 (a), 6 (b) and 9 (c) weeks after sowing in Kokeyai and Samjan experimental sites. The error bar indicates SE for the mean for each crop and seed-placement depth.

Discussion

Water status of crops

Although during the growing season, there was very little rainfall, soil water content of both locations remained in the plant-available ranges throughout the soil profile even though in the late growing season top soil layers dried to close to wilting point. Among the legumes, peanut in general has the highest water use, followed by cowpea, soybean and mungbean (Pandey et al. 1984a). Lawn (1982) suggested that cowpea and mungbean did not extract soil-water as rapidly as soybean. Previous studies by Kerdsuk (1986) found that peanut after rice in the dry season without irrigation required 474 mm of water for growth, which exceeded the sum of rainfall during the growing season plus the amount of stored soil water in the upper 1 m of the soil profile at planting (295 mm). In addition to the residual soil moisture left over from the preceding rainy season, capillary water rise from a shallow water table was the main factor controlling soil water supply and hence growth and yield of peanut after rice (Kerdsuk 1986). There were no leaf wilt symptoms and crops had satisfactory growth in both locations suggesting that in the present study the access to water from the shallow water table was sufficient to maintain water supply throughout the growing season.

The shallow water table, which decreased to 130 cm depth at Kokeyai and 103 cm at Samjan, supplied enough water for growth so yield of shallow- and deep-placed plants did not differ significantly. However, in locations with a deeper water table, growth and yield of deep-placed plants in previous studies were markedly higher than shallow-placed and attributed to less soil water content in the main root zone of shallow-placed plants (Jintrawet et al. 1983, Kerdsuk 1986, Polthanee 2001). Normal (shallow) seed-placement was more suitable than deep seed-placement for growing legumes in the dry season without irrigation when there was a shallow water table and water content was above PWP throughout soil profiles in the growing season. Kerdsuk (1986) reported that sites most suitable for growing crops in the dry season without irrigation should have a water table depth between 30–50 cm at planting and not more than 150 cm at harvesting. Thus both the present sites fitted these criteria except that the initial water table depth at Kokeyai was lower than optimal.

The limited rainfall during the growing season could be beneficial to growth of legumes sown with deep-placed seed. However, Jintrawet et al. (1983) and Polthanee (1991) suggested that a lot of rainfall in the dry season, especially during the first 4 weeks after seed-placement, was detrimental as it led to soil compaction and poor emergence of deep-placed seeds of legumes. Rao and Reddy (1985) reported that peanut seeded 10 cm deep in soil had poor development of root and shoot at the seedling stage when seeds were covered with packed soil. However, a lot of rainfall might be an advantage for shallow-placed legumes by replenishing topsoil water where most roots of shallow-placed legumes occurred.

Rooting patterns of cowpea, mungbean, peanut and soybean at different seed-placement depth

The rooting patterns of cowpea, mungbean and peanut in this experiment were not markedly different perhaps because the experiment had enough water for legume growth throughout the growing season even in Kokeyai that had the lower soil water content. There was also no clear difference in AWR among legume species or between locations.

Unlike cowpea, mungbean and peanut, soybean roots were less responsive to seed-placement depth: at different seed-placement depths, soybean had no significant change in total root dry weight and length density. While root distribution of cowpea, mungbean and peanut after deep seed-placement was deeper than those legumes after shallow seed-placement, particularly late in the growing season and in the drier location, rooting depth of soybean with different seed-placement depths was not different. Legume roots in both locations extended in depth over time following the declining water table. Nevertheless, depth of roots in Samjan was shallower than Kokeyai because Samjan had shallower water table. However, with deep seed-placement at Samjan roots of cowpea and peanut especially penetrate into the water table whereas at Kokeyai roots remained above the water table. Excessive moisture from the shallow water table in Samjan might inhibit root growth at depth by hypoxic stress and trigger aerenchyma development in roots (see below). On account of their deeper roots, cowpea, mungbean and peanut were likely to be accessing stored water from the water table to a greater extent than soybean.

With both shallow and deep seed-placement, soybean roots were mostly distributed in the topsoil at both the drier and wetter sites. Peanut, cowpea and mungbean after deep seed-placement had greater root depth but reduced root mass. By contrast those legumes sown with shallow seed-placement had more root mass especially at 0–15 cm but maximum root depth was less.

Roots of cowpea, mungbean and peanut that could respond to seed-placement depths and soil moisture, facilitate better growth and survival in dry seasons without irrigation when the plants depend mainly on capillary water rising from a gradually dropping water table. The higher root distribution at greater soil depths enhanced drought tolerance and thus could stabilize pod yield and harvest index under water-stress conditions (Songsri et al. 2008). Barraclough et al. (1989) reported that little root growth of winter wheat occurred in the topsoil during the drought but there was compensatory growth in the subsoil provided that nitrogen fertilizer was given.

Limited rooting very likely decrease the efficiency of water and fertilizer use by soybean (Robertson et al. 1980). Benjamin and Nielsen (2006) indicated that water deficit did not affect the relative soybean root distribution but water deficit resulted in a greater proportion of chickpea and field pea root growth deeper in the soil.

The present results explain those of Khongmorn (1994) who demonstrated that cowpea and peanut were the crops most suitable for planting after rice when reliant on capillary water rise from the water table in north-east Thailand. Grain yield of cowpea and peanut decreased only slightly when grown in the dry season relying on residual soil moisture while soybean yield decreased significantly in such conditions. Pandey et al. (1984b) indicated that cowpea and peanut had greater root densities at 0.4 to 0.8 m, particularly in the lowest soil moisture regime, but soybean and mungbean had relatively shallow root systems and did not maintain water uptake in the lowest soil moisture regime. Mungbean did not behave the same as soybean in the present study, as expected from previous research (Pandey et al. 1984b), perhaps because legumes in the present study did not experience water deficit.

Growth duration of crops should also be a consideration with deep seed-placement because it determines the total water requirement of each crop. Four legumes had different harvesting periods, ranging from 10 WAS for cowpea (Suddhiyam and Kowsurat 2001) and mungbean (Thanomsub and Watanasit 2001), 10–11 WAS for soybean (Srisombun and Kaewmeechai 2001) and 12–15 WAS for peanut (Sarawat 2001). If water stress happened at a later stage of crop growth (after 9 WAS) due to a rapidly dropping water table, peanut would encounter water deficit but cowpea, mungbean and soybean would escape it. However the avoidance of water stress in peanut growth depended on the ability of roots to extend to the wetting front of receding soil water. Clearly the increased density of peanut roots at the deeper soil layer under more severe stress conditions would contribute to drought avoidance and to yield (Kashiwagi et al. 2006).

Seed-placement depth and rooting patterns of legumes

Deep seed-placement generally decreased overall root weight and root length. However, seed-placement depth markedly altered the distribution by depth of the roots of the four legume species. After deep seed-placement, roots of peanut, cowpea, mungbean and to a lesser extent soybean, were relatively more abundant in subsoil (more than 15 cm below the soil surface) while roots of these legumes after shallow seed-placement were mainly distributed in the topsoil (0–15 cm from soil surface). This response can be attributed to the higher soil water content in the topsoil of shallow-placed treatment than deep-placed treatment since soil water contents can affect root depth and distribution (Pandey et al. 1984b). There were a few roots in the topsoil of deep-placed legumes perhaps because of seed-placement method. Deep-placed seeds were accomplished using deep ploughing that interrupted capillary rise and disturbed the topsoil causing a lower topsoil water content than with shallow sowing. Subsequently it appears that growth of topsoil roots of deep-placed legumes was limited by the drier topsoil.

Dry topsoil of deep-placed legumes could drive legume roots to penetrate to deeper soil layers. On the other hand with wet topsoil of shallow-placed legumes, roots did not penetrate deeper in search of water. Polthanee (2001) reported that peanut after deep seed-placement (15 cm) had a higher root length density at the 30–45 cm soil depth than after shallow seed-placement (5 cm) and this could facilitate greater growth and yield. Yusuf Ali et al. (2005) also found that greatest chickpea (Cicer arietinum L.) root proliferation was in the soil surface in soil with irrigation but in soil layers down to 90 cm depth in soil without irrigation because of greater depletion of available water in surface soil. Moreover, Krishnamurthy et al. (1998) reported that root length to weight ratio of chickpea declined gradually in the 0–10 cm soil depth but showed an apparent increasing trend at 30–135 cm soil depths under receding soil moisture conditions due to death and decay of fine and newly formed laterals as a result of drying of the soil surface. However, in both seed-placement depths, the roots emerged densely under and near the seed placement position. So, there were only a few roots in the soil layer over the seed position, 0–5 cm soil layer for shallow-placed legumes and 0–15 cm layer for deep-placed legumes.

By increasing the root surface area to root weight ratio (AWR), deep-placed legumes in the present study would be more efficient in water uptake than shallow seed-placement legumes. Higher AWR may mean relatively more fine roots in the root system that should lead to better soil exploration and water extraction (Benjamin and Nielsen 2006). Alternatively, increased aerenchyma in roots would decrease root weight per unit length that could also increase AWR of deep-sown legumes. After deep seed-placement, roots penetrated deeper into soil layers closer to the water table with lower levels of oxygen supply especially at Samjan. Legume roots might adapt to hypoxic conditions by producing aerenchyma (Drew 1997). Aerenchyma has also been suggested to decrease the energy cost of root production and decrease the P requirements for root elongation which would be useful in low P subsoils such as in the present deep sands (Fan et al. 2003).

Therefore, deep seed-placement promoted root growth of peanut, cowpea and mungbean into the subsoil that stored more water in the dry season. Moreover, with deep seed-placement roots of these legumes penetrated deeper in the soil to exploit subsoil water. Roots of deep-placed legumes also had higher AWR which may arise from finer roots or increased aerenchyma, both of which could lead to better soil exploration and water extraction. Unlike peanut, cowpea and mungbean, soybean rooting depth did not respond much to seed-placement depths, soil moisture and time. Soybean was not suitable for dry season deep seed-placement on sandy soils without irrigation condition because its roots could not exploit deep soil moisture as well as other species even after deep seed-placement. However, deep seed-placement may not be necessary if soil water is available at shallow depth for legumes.

Acknowledgements

The authors are grateful for the financial support of the Royal Golden Jubilee PhD Program (Grant no. PHD/0268/2545) under the Thailand Research Fund and also for support by the Farming System Research Group of KhonKaen University.