Review of wheat improvement for waterlogging tolerance in Australia and India: the importance of anaerobiosis and element toxicities associated with different soils

Abstract

Background and Aims

The lack of knowledge about key traits in field environments is a major constraint to germplasm improvement and crop management because waterlogging-prone environments are highly diverse and complex, and the mechanisms of tolerance to waterlogging include a large range of traits. A model is proposed that waterlogging tolerance is a product of tolerance to anaerobiosis and high microelement concentrations. This is further evaluated with the aim of prioritizing traits required for waterlogging tolerance of wheat in the field.

Methods

Waterlogging tolerance mechanisms of wheat are evaluated in a range of diverse environments through a review of past research in Australia and India; this includes selected soils and plant data, including plant growth under waterlogged and drained conditions in different environments. Measurements focus on changes in redox potential and concentrations of diverse elements in soils and plants during waterlogging.

Key Results

(a) Waterlogging tolerance of wheat in one location often does not relate to another, and (b) element toxicities are often a major constraint in waterlogged environments. Important element toxicities in different soils during waterlogging include Mn, Fe, Na, Al and B. This is the first time that Al and B toxicities have been indicated for wheat in waterlogged soils in India. These results support and extend the well-known interactions of salinity/Na and waterlogging/hypoxia tolerance.

Conclusions

Diverse element toxicities (or deficiencies) that are exacerbated during waterlogging are proposed as a major reason why waterlogging tolerance at one site is often not replicated at another. Recommendations for germplasm improvement for waterlogging tolerance include use of inductively coupled plasma analyses of soils and plants.

‘No grain is ever produced without water, but too much water tends to spoil the grain and inundation is as injurious to growth as dearth of water.’ Narada Smriti XI, 19; circa 3000 bc.

‘Waterlogging’ is defined as a condition of the soil where excess water limits gas diffusion; while ‘waterlogging tolerance’ is defined as survival or the maintenance of high growth rates, biomass accumulation or grain yield under waterlogging relative to non waterlogged (usually drained soil) conditions (Setter and Waters, 2003).

INTRODUCTION

Considerable research has been conducted on waterlogging tolerance of crops, yet little of this has involved controlled conditions in diverse environments, diverse soils, or with large germplasm sets. This lack of information under these conditions is probably not surprising because there is evidence that waterlogging-prone environments have considerable diversity and complexity. Diversity occurs in the timing, duration and severity of waterlogging stress (Setter and Waters, 2003) as well as the physical and chemical changes associated with waterlogged soils (Ponnamperuma, 1972, 1984; Marschner, 1984; Kirk et al., 2003); other diversity exists for abiotic factors associated with waterlogging-prone environments (Drew and Lynch, 1980; Laanbroek, 1990). These characteristics have led to many questions relevant to germplasm improvement which continue to be debated today, e.g. ‘Whether the principal cause of damage to the shoots of plants grown in waterlogged soil arises more from the production of toxic substances in the soil than from the oxygen deficiency experienced by roots has long been questioned’ (Drew, 1981).

The complexity of the waterlogging-prone environment would lead one to predict that waterlogging tolerance may be difficult to repeat in different soils or environments. For example, the wheat variety Ducula-4 was found to have one of the highest levels of tolerance after waterlogging for over 3 months in a soil at Obregon, Mexico (vanGinkel et al., 1992; Sayre et al., 1994). However, in Australia (Condon, 1999a, b; Setter, 2000) and in India (pers. obs.) it has not been possible to confirm the waterlogging tolerance of Ducula-4, with many local varieties out-performing this variety in drained and waterlogged conditions. In fact in some soils (Kaithal, India), Ducula-4 is one of the most waterlogging intolerant varieties relative to plants grown in drained soil (pers. obs.). When this genotype was first screened in Mexico it was probably in an environment free of additional constraints; growth was at high nutrition (up to 200 kg N ha⁻¹; vanGinkel et al., 1994), high temperature, and water was available after waterlogging treatments in contrast to a dry finish in environments like Australia. Such waterlogging conditions in Mexico are clearly different from those in Australia and India.

McDonald et al. (2006) screened 17 wheat varieties in 11 different field locations in Western Australia with low, moderate or severe waterlogging under natural conditions, where grain yield was reduced by approx. 10, 25 and 65 %, respectively. What is particularly striking about this work is that, a variety that was one of the most waterlogging-tolerant varieties in one trial, was just as likely to be one of the most intolerant varieties in another trial. For example, the Australian variety Cascades was one of the best performing varieties under waterlogging in three trials with little or no yield reductions, and it was one of the worst performing varieties in three different trials with up to 73 % yield reductions relative to non (or less) waterlogged plants. A major constraint of waterlogging trials in natural rainfed environments is that the timing, duration and severity of waterlogging is often different in different trials, therefore making the varietal comparisons impossible to interpret, i.e. a variety may be tolerant to waterlogging at the germination and emergence stages, but it may not be tolerant at the vegetative or reproductive stages (Setter and Waters, 2003).

The diversity and complexity of waterlogging-prone environments also suggest that for each set of conditions, there could be different priorities for mechanisms of tolerance. There are extensive reviews on mechanisms of tolerance to waterlogging or flooding (Jackson and Drew, 1984; Armstrong et al., 1994) or specifically anoxia (Greenway and Gibbs, 2003; Gibbs and Greenway, 2003); with Setter and Waters (2003) presenting 22 mechanisms of tolerance spanning across phenology, morphology, nutrition (including deficiencies and toxicities), root metabolism, recovery and prevention of post-anoxic damage. The prioritization of so many mechanisms of tolerance to waterlogging can seem daunting, particularly for germplasm improvement programmes with massive germplasm sets and limited resources for screening. However, many of these mechanisms can be grouped according to the relevance for different stages of plant development, therefore providing a preliminary basis for prioritization.

Recent analyses of waterlogging-prone soils in Australia and India have indicated several factors that may contribute to the diverse effects of waterlogging on plant growth in different environments. N. P. S. Yaduvanshi et al. (unpubl. res.) evaluated changes in redox potentials and soil manganese (Mn) and iron (Fe) during waterlogging in four different soils. The Australian and Indian soils showed similar decreases in redox potentials with time after waterlogging at approx. 20 °C. However, as Australian soils are naturally waterlogged during winter at about 5–10 °C, there are generally much slower reductions in redox potentials in Australian than in Indian soils in the field. One of the most significant differences between the Australian and Indian soils is associated with changes in Mn and Fe during waterlogging; concentrations of DTPA-Mn and DTPA-Fe increased above critical toxicity concentrations in all soils. Two Indian soils tended to be 2–10 times higher in DTPA-Mn than three Australian soils; and Australian soils were up to 10 times higher in DTPA-Fe than the Indian soils. These increases were up to 10 and 60 times higher, respectively, than reported critical concentrations for wheat (N. P. S. Yaduvanshi et al., unpubl. res.).

Microelement toxicities, predicted from soil analyses during waterlogging, have been substantiated by plant analyses for wheat grown in waterlogged soils in the USA (Ding and Musgrave, 1995; Musgrave and Ding, 1998) and Australia (Khabaz-Saberi et al., 2006). Ding and Musgrave (1995) found that Fe, Mn, phosphorus (P), potassium (K), sodium (Na) and sulfur (S) were affected in wheat grown in waterlogged river silt, and they subsequently used increases in the sum of Fe, Mn and P concentration in root samples as an indicator of waterlogging intolerance (r² = 0·94; Musgrave and Ding, 1998). However, this correlation was taken across both drained and waterlogged treatments which showed clear clusters, while the relationship within waterlogging treatments alone was much lower for both the sum of elements (Musgrave and Ding, 1998) as well as when individual elements were evaluated (Ding and Musgrave, 1995).

Microelement increases were measured in wheat shoots of six wheat varieties during 49 d waterlogging in an acidic and neutral soil from Australia and in potting mix (Khabaz-Saberi et al., 2006). These data showed that the ranking of varieties changed in different soils, even though soils were all waterlogged at the same time (3 weeks after germination), the same duration (49 d), and identical climatic (temperature, irradiance and humidity) conditions. For plants grown in waterlogged acidic soil, shoot concentrations of Al, Mn and Fe increased by 2- to 10-fold relative to plants grown in drained soil, and they were above critical concentrations for toxicities.

One well-known interaction with soil elements during waterlogging occurs for salinity. There is clearly an interaction of salinity and anaerobiosis affecting Na concentrations in the shoots of wheat; shoot Na concentrations are approx. 2 times higher when plants are waterlogged relative to drained sand-culture conditions (Barrett-Lennard, 1986; see also review by Barrett-Lennard, 2003a). Among non-halophytes, tolerance to a combination of salinity and waterlogging is suggested to be greater in a waterlogging-tolerant plant like rice than in a waterlogging-intolerant plant like barley (Barrett-Lennard, 1986). However, these conclusions are based on published data where there are different treatment durations; comparisons of waterlogged sand (barley) versus solutions bubbled with nitrogen (rice); and different salt concentrations (0·5–125 and 2–80 mm for barley and rice, respectively). There are no published data clearly demonstrating salinity–waterlogging interactions affecting the growth of cereals, even though visual evidence for this is convincing (photo 3·5 of Barrett-Lennard, 2003b). Previously unpublished data are therefore presented here in support of evidence for such important interactions between waterlogging/hypoxia and salinity/NaCl (Discussion).

Setter et al. (2004) proposed that waterlogging tolerance is often a product of tolerance to anaerobiosis and to element toxicities, e.g. Mn, Fe, Na, Al and B, in different soils (Discussion). This hypothesis is based on the work of Barrett-Lennard (2003a), Khabaz-Saberi et al. (2006), McDonald et al. (2006) and other unpublished observations for wheat grown under anaerobic conditions and in waterlogging-prone environments in Australia and India. Experiments presented here provide further evidence in support of this hypothesis and highlight the complex diversity of waterlogging-prone environments.

MATERIALS AND METHODS

Germplasm used in screening trials in Australia and India included 37 wheat varieties (Triticum aestivum L), Hordeum marinum, and a doubled haploid population Ducula-4/2 * Brookton containing 197 genotypes.

Methods used in Australia

In Australia, plants were sown in 4-L pots in the target environment in the field at the approximate time of the field season (May/June each year) to assure the same temperature and irradiance conditions of plants grown under natural conditions (Fig. 1).

Fig. 1.

Waterlogging tolerance screening ponds at Katanning, Western Australia. Separate ponds enable screening of genetically fixed or segregating populations under controlled conditions. In this case up to 10 000 pots were used to screen approx. 450 genotypes in four soils. The insert shows normal (left) and specially designed pots (right) to minimize root escape during waterlogging treatments (see Materials and Methods).

Pre-germinated seeds were sown in each pot (five plants per pot) at a depth of 2 cm with four replicates (pots) per genotype. Ten Western Australian soils or potting mix were used; soils were topsoils or subsoils from waterlogging-prone regions in Western Australia (Table 1). All these soils are sandy duplex soils except Mindarabin which is a grey clay; soil pH and selected elements in these soils, with and without waterlogging, are presented in Table 1. Microelements were determined based on exchangeable (Exch) or diethylenetriaminepentaacetic acid (DTPA)-extracted soils using inductively coupled plasma (ICP). DTPA-extracted soils enable microelement concentrations to be analysed in the same form/concentration as they exist during waterlogging. Melich No. 3 extractions are used for some microelements since this can be done in one soil extract, it highly correlates to DTPA extracts (r² = 0·9; H. Khabaz-Saberi, unpubl. res.; cf. Allen and Walton, 2003), and reduces analyses cost and sample size. The minimum critical concentrations for toxicity in soils stated in Table 1 are based on published values for electrical conductivity (EC) (Maas and Hoffman, 1977; Shaw, 2005), boron (B) (Ponnamperuma et al., 1979; Bell, 2005), aluminium (Al) (Slattery et al., 2005), Mn (Slattery et al., 1995; Uren, 2005) and Fe (>30 mg kg⁻¹: McFarlane, 2005; >35 mg kg⁻¹: Malewar and Ismail, 1995).

Table 1.

Soil physical and chemical characteristics in drained or waterlogging treatments

Soil	Location	pH drained (CaCl2)	pH WL (CaCl2)	OC ( %)	EC (1 : 5) (mS m−1)	Na drained (Melich) (mg kg−1)	B drained (Melich) (mg kg−1)	Al drained (Exch) (mg kg−1)	Mn drained (DTPA) (mg kg−1)	Mn WL (DTPA) (mg kg−1)	Fe drained (Melich) (mg kg−1)	Fe WL (Melich) (mg kg−1)	Soil type (sand : silt : clay)
Mindarabin (Wemyss site), WA	33°52'04·3′′S, 118°21′03·9′′E	6·3	7·4	0·73	43	140	2·5	<1·8	7·6	9·4	66	120	Grey clay (66 : 5 : 29)
Potting mix	–	5·3	5·7	7·86	46	110	0·2	<1·9	4·1	7·3	140	180	(97 : 3 : 0)
Holly Siding, WA	33°48′43·7′′S, 117°27′49·1′′E	4·6	5·6	1·02	25	4	0·1	17·1	5·9	18	230	510	Sandy duplex (84 : 7 : 9)
Katanning (0–30 cm), WA	33°40′57·4′′S, 117°37′45·7′′E	4·7	6·0	0·63	20	<1	<0·1	11·7	5·1	15	76	190	Sandy duplex (94 : 4 : 2)
Topsoil (0–10 cm)		5·0	5·6	0·79	20	1	0·2	3·6	7·3	6·5	94	120	–(94 : 4 : 2)
Subsoil (20–30 cm)		4·4	5·9	0·13	19	<1	<0·1	4·5	2·1	10	47	190	–(98 : 2 : 0)
Mt Barker, WA	34°34′56·1′′S, 117°30′20·2′′E	5·0	5·6	2·34	40	10	0·3	15·3	8·8	19	170	240	Sandy duplex (90 : 6 : 4)
South Stirlings (Warburton 2004 site), WA	34°32′51·7′′S, 118°05′15·7′′E	4·8	6·2	1·67	(–)	13	0·2	7·2	2	1	160	120	Sandy duplex (–)
Esperance, WA	33°37′08·4′′S, 121°47′02·4′′E	4·4	5·6	1·10	22	1	<0·1	13·5	5·3	3·6	94	88	Sandy duplex (95 : 3 : 2)
South Stirlings (Howard 2003 site), WA	34°37′28·3′′S, 118°09′19·6′′E	5·3	5·6	0·80	14	2	0·1	1·8	1·8	0·6	420	470	Sandy duplex (94 : 4 : 2)
NDUAT, U.P., India	26°32′37·8′′N, 81°49′26·9′′E	9·0 (H20)	(–)	0·14	152–180	1150–2070	0·6–2·1	(–)	19·9	21·6	7	8	Alkali silty loam (47 : 37 : 16)
CSSRI, Haryana, India	29°42′22·8′′N 76°57′10·6′′E	8·5 (H20)	(–)	0·35	100	(–)	0·5	(–)	4	48	7	36	Alkaline silty clay (–)
CSSRI, Haryana, India		9·2 (H20)	(–)	0·26	260	(–)	0·5	(–)	3	57	4	19	Alkaline silty clay
Critical conc. (toxicity)					>60–100		>3	>2–4	>10–20	(–)	>30–35	(–)

Soils were waterlogged for 49 d (Australia) or 14 to 21 d (India). Where there were no substantial differences in elements between drained and waterlogged (WL) treatments, e.g. for Na, B and Al, only values for drained treatments are presented.

Critical concentrations for toxicities are as described in the text.

Melich, Melich No. 3; DTPA, diethylenetriaminepentaacetic acid; OC, organic carbon; WA, Western, Australia.

Pots (holding approx. 4 kg soil; 25 cm high) were specially designed to minimize root escape by (a) placing drainage holes in a raised-up section at the bottom of the pot in the centre rather than the sides (see insert to Fig. 1); and (b) lining the inside of each pot with a piece of root mat (Source: Network Trading Co., NSW, Australia). Each pot was fertilized with optimum concentrations of nutrients at the following rates (μmol g⁻¹ soil): NH₄NO₃, 4·46; KH₂PO₄, 0·44; K₂SO₄, 0·92; CaCl₂.2H₂O, MgSO₄.7H₂O, 0·09; ZnSO₄.7H₂O, 0·04; MnSO₄.H₂O, 0·07; CuSO₄.5H₂O, 0·024; H₃BO₃, 0·013; Na₂MoO₄.2H₂O, 0·008. In preliminary experiments plants were fertilized with 0·5, 1, 2 and 4 times these nutrient concentrations and the selected levels shown here provided the optimum growth under drained conditions for all Australian soils (data not presented). All pots were watered to field capacity every second day and grown under drained conditions until 21 d after sowing, i.e. when leaf 3 was half emerged.

At 21 d after sowing, four replicate pots were used per variety for each waterlogged and drained treatment. For waterlogging treatments, one pot was placed into each of four replicate ponds (approx. 15 × 25 m) constructed in the field. Ponds were laser levelled and lined with a double layer of polyethylene plastic (200 µm thickness) to provide water depths to the soil surface of pots. Pots were completely randomized within each pond or drained blocks. After waterlogging treatments, shoot samples were collected, dried at 70 °C for 48 h, weighed and then sent to the Waite Analytical Services, University of Adelaide for element analyses by ICP. The minimum values for critical concentrations for element toxicities in wheat shoots or leaves as shown in the Results (dashed lines in Figs 5 and 6) have been published for Na (>0·8 %; Weir and Cresswell, 1994 as cited by Reuter et al., 1997), Al (>50 mg kg⁻¹; Ma et al., 2003), B (>10–20 mg kg⁻¹, Mortvedt, 1972; Ascher-Ellis et al., 2001), Mn (>100 mg kg⁻¹; Singh and Rao, 1995; Reuter et al., 1997) and Fe (>100 mg kg⁻¹; Reuter et al., 1997). Possible contamination of plant samples with soil was evaluated here by measurements of titanium in all samples (Cherney and Robinson, 1983; Cary et al., 1986; L. Palmer, pers. comm.). This was validated in preliminary experiments where soil (5, 10, 25 and 50 mg) was added to 1-g plant samples, i.e. at 0·5–5 % w/w. Any plant samples with titanium equivalent to 1 % or more soil contamination were deleted from the mean data for other elements.

Fig. 5.

Concentrations (mg kg⁻¹ ± s.e.m.) of (A) shoot Mn, (B) shoot Fe and (C) shoot Al of three wheat varieties grown under drained and waterlogged treatments in Katanning, Esperance and South Stirlings (Warburton) soil or in potting mix. Plants were grown in waterlogged soil for 49 d. Dashed lines indicate critical concentrations for toxicity (see Materials and Methods).

Fig. 6.

Concentrations (mg kg⁻¹ ± s.e.m.) of (A) shoot Al, (B) shoot B, and (C) leaf Na ( % ± s.e.m.) of wheat varieties exposed to 12 d under drained and waterlogged treatments in Karnal soil at pH 7·8–8·2 or 9·2–9·4 (Karnal, India). Plants in (C) were in a different experiment from (A) and (B) so the soil pH was slightly different. Dashed lines indicate critical concentrations for toxicity (see Materials and methods).

Varieties were analysed for Al tolerance at DAFWA, South Perth, WA, by measurements of maximum root length. Varieties were exposed with or without 440uM AlCl₃.6H₂0 (12 ppm) in aerated solution culture for 10 d at pH 4·1–4·2 (Raman et al., 2005). Plants were completely randomized, and data presented are means of three replicates of five plants.

Methods used in India

In India, waterlogging-tolerance trails were conducted with 10-L pots using alkaline sodic soil from the Central Soil Salinity Research Institute (CSSRI), Karnal, Haryana, India (Table 1). Five plants per pot were sown at the approximate time of the field season (November each year) to assure relevant temperature and irradiance conditions of plants grown under natural conditions. Plants were grown in a screenhouse until approx. 3 weeks when waterlogging treatments were imposed by saturating soils in each pot with water to approx 0–1 cm above the soil surface. Four replicate pots were used for each genotype for waterlogged and drained treatments.

Field plants in India were grown at the CSSRI Field Station in the same alkaline sodic soils as used for pots (Singh et al., 2006); or in alkaline sodic soils at the Narendra Deva University of Agriculture and Technology Field Station (NDUAT), Kumarganj, U.P. India (Table 1). Genotypes were evaluated in randomized block design with three replications per genotype. Plants were waterlogged for 10–15 d at 22 d after sowing at CSSRI and for 7–10 d at 22 d after sowing at NDUAT. These different times of waterlogging treatment were necessary to optimize discrimination between genotypes in these different field locations.

Redox potential measurements were made of soils in Australia and India using redox electrodes as described by Setter and Waters (2003) and N. P. S. Yaduvanshi et al. (unpubl. res.); and data from different soils were all standardized to pH 7 (Patrick et al., 1996). Electrodes were inserted vertically into soils, and measurements were taken at a depth of 5 cm. ICP analyses were made of whole intact plant shoots following nitric acid digestion in India at Central Research Institute for Dry Land Agriculture (CRIDA), Hyderabad, Andhra Pradesh, India.

RESULTS

Screening method development

Development of the waterlogging-tolerance screening method, using plants grown in pots which were waterlogged in ponds, was conducted in Western Australia for three seasons before a robust, reliable and efficient screening method was developed. This occurred because (a) there are large differences in waterlogging tolerance for different soils (this was not known; see below); (b) soils needed to be thoroughly mixed before filling pots so as to minimize variation within replicates; and (c) root escape from the bottom of pots reduced the adverse effects of waterlogging on plant growth and resulted in extremely variable shoot dry weights (data not presented; Fig. 1). The Australian soils used here were all selected from waterlogging-prone environments with low levels of salinity; however, they each had unique characteristics associated with pH, organic matter, and concentrations of other elements like Na, B, Al, Mn and Fe (Table 1).

Waterlogging tolerance and correlations between different environments or different soils

Typical shoot dry weights at the end of waterlogging or drainage treatments in an Australian soil are shown in Fig. 2 where 38 genotypes are ranked based on waterlogging tolerance, i.e. the relative weights of waterlogged and drained plants. This highlights that there is about a 2-fold difference in both biomass at the end of 49 d waterlogging and waterlogging tolerance for plants waterlogged in this soil and under these conditions (calculated from Fig. 2).

Fig. 2.

Varietal tolerance of wheat to waterlogging in Katanning soil. Varieties were waterlogged for 49 d at 21 d after sowing. Varieties are ranked according to ‘waterlogging tolerance,’ values, i.e. waterlogged/drained shoot dry weights given in parenthesis. Data are the same as for Table 2 (2004).

Reproducibility of the response to waterlogging in different experiments was measured by conducting two experiments in the same year, with the same genotypes, the same seed, the same soil, and same waterlogging treatments. This shows that there is a high correlation (r² = 0·94) for shoot dry weight of plants of ten genotypes grown in drained and waterlogged conditions for 49 d (Fig. 3). There is an obvious clustering of points in these drained and waterlogged treatments which also occurs in other experiments. However, if waterlogging duration is reduced relative to that used here, it is likely that the points would have formed a continuous relationship. The r² value for shoot dry weights in the waterlogging treatments alone for these two experiments is 0·56 (calculated from Fig. 3).

Fig. 3.

Comparison of shoot dry weight per plant (DW/pl) in different experiments (A and B) for ten wheat genotypes at the end of 49 d waterlogging or drained treatments. The r² value is for all data; the r² value for waterlogged treatments only in these two experiments is 0·56.

When waterlogging tolerance is compared based on randomly paired replicates of waterlogged and drained treatments, the correlations are reduced (r² = 0·31) due to addition of variation resulting from drained plus waterlogged treatments together. In other experiments comparing the response to waterlogging for up to 16 varieties screened in different years, there is also a high correlation (r² = 0·73) based on biomass at the end of waterlogging, as long as screening is conducted under the same conditions and in the same soil.

A two-way analysis of variance of data in Fig. 2 is shown in Table 2. For the combination of drained and waterlogged treatments, treatment effects explain 78 % of the variance and variety effects explain 16 % of the variance. Such results could be expected based on the severity of waterlogging treatments used here where there is a 50–75 % reduction in biomass. The variety × treatment effect is also significant (P < 0·001) which means that not all varieties performed in the same way, i.e. some are reduced more by waterlogging than others. However, only a small portion of the variation (4 %) is explained by this interaction, and it does not account for large effects overall.

Table 2.

Analysis of variance of waterlogging tolerance data based on drained and waterlogged treatments in Fig. 1

Source of variation	d.f.	% variance	Mean square	Variance ratio	F probability
Rep stratum	3	0·1	0·03418	1·16
Variety	37	15·9	0·29433	10	<0·001
Treatment	1	77·5	53·09783	1804·32	<0·001
Variety × treatment	37	4·4	0·08057	2·74	<0·001
Residual	208	9·0	0·02943

In different soils, the correlations (r²) of 32 genotypes for waterlogging tolerance in three different Australian soils and potting mix were usually <0·1 (Table 3A). In these experiments plants were grown in pots and waterlogged at the same time (21 d after sowing) for the same duration (49 d) under identical climatic conditions. The correlations between different soils increased to 0·1–0·3 when measurements were based only on biomass at the end of waterlogging (Table 3B) rather than ‘waterlogging tolerance’ (Table 3A). Poor correlations were also observed between different soils with a doubled haploid population with 192 lines grown in South Stirlings (Warburton), Esperance and Katanning soils. Doubled haploid lines waterlogging for 49 d were reduced in biomass by 0–86 % in different lines; however, the correlations for waterlogging tolerance or for biomass at the end of waterlogging between any two soils ranged from only 0·03 to 0·06 and 0·00 to 0·09, respectively.

Table 3.

Correlation tables of (A) waterlogging tolerance and (B) biomass at the end of waterlogging

Soil	Katanning	South Stirlings (Warburton)	Esperance	Potting mix
(A) Waterlogging tolerance (biomassWL/biomassDrained)
Katanning	–	0·034	0·001	0·045
South Stirlings (Warburton)	–	–	0·131	0·015
Esperance	–	–	–	0·001
Potting mix	–	–	–	–
(B) Biomass at end of waterlogging
Katanning	–	0·20	0·20	0·32
South Stirlings (Warburton)	–	–	0·22	0·10
Esperance	–	–	–	0·15
Potting mix	–	–	–	–

Plants were grown under drained conditions for 21 d and then treated under waterlogging or drained conditions for 49 d. Data are based on 32 wheat lines (two isogenic lines, 23 Australian varieties, two doubled haploid lines and five Indian varieties) grown in four soils in pots at Katanning, WA, in 2006.

Waterlogging tolerance screening trials in India similarly showed little or no correlation between four different locations/soils based on biomass at the end of waterlogging, grain yields of plants exposed to waterlogging, or waterlogging tolerance calculated from either (relative) biomass or grain yields. These field trials were conducted on research stations with soil pH ranging from 7·2 to 10·4 across the wheatbelt region of northern India. When grain yields were compared for 187 Ducula-4/2 × Brookton doubled haploid lines exposed to waterlogging for 10–12 d, there was an approx. 3-fold range in grain yields at each of the four field locations; however, the correlations (r²) between any two out of four field locations ranged from only 0·00–0·02 (data not presented).

In order to more fully evaluate the effects of different soils on waterlogging tolerance, ten soils were compared under identical temperature and irradiance conditions, with identical timing and duration of waterlogging, using plants grown in pots exposed to drained and waterlogged conditions in the field at Katanning, Western Australia. The response of these two varieties, grown in waterlogged soil for 49 d ranged from 82 % reductions in biomass at the end of waterlogging in the acidic South Stirlings (Howards) soil to a significant improvement in growth in a waterlogged acidic subsoil from Katanning (Table 4). In this experiment, the correlation between shoot and root dry weights for these ten different soils and potting mix was (r² = 0·94) for both varieties. In comparison, in the analysis of 38 varieties in Fig 2, the correlation between shoot and root dry weights after waterlogging in one soil had an r² = 0·66.

Table 4.

Waterlogging tolerance for wheat varieties Westonia and Camm in ten different soils

Soil type	Westonia	Camm
Mindarabin (Wemyss site)	0·92	1·07
Potting mix	0·87	0·84
Holly Siding	0·70	0·90
Katanning (0–30 cm)	0·62	0·41
Topsoil (0–10 cm)	0·65	0·84
Subsoil (20–30 cm)	1·12	1·31
Mt Barker	0·62	0·70
South Stirlings (Warburton 2004 site)	0·47	0·22
Esperance	0·32	0·29
South Stirlings (Howard 2003 site)	0·29	0·18

Waterlogging tolerance was determined after 49 d waterlogging and is based on biomass in waterlogged relative to drained treatments for plants grown in pots in the field at Katanning, WA.

All soils, including potting mix, were reduced in redox potentials during waterlogging with most soils being below the 350 mV level associated with anoxia (Marschner, 1984) within 5–15 d (Fig. 4). At the end of waterlogging treatments, soils had fallen to redox potentials of –25 to –225 mV (Fig. 4). In contrast, drained soils remained at redox potentials of 500–600 mV throughout the experimental period (data not presented). At the end of 49 d of waterlogging treatments, soils were drained, and the redox potentials slowly increased except when there was rain during 52–55 d after waterlogging. Similar changes in redox potentials during waterlogging were observed in another year with four soils including potting mix.

Fig. 4.

Soil redox potentials of eight soils during waterlogging in the field at Katanning, Western Australia. Soils are the same as used in Tables 1 and 4; some soils with intermediate values are not shown for clarity. Data are standardized to pH 7 (see Materials and Methods); and soils are considered anoxic at redox potentials of ≤350 mV at pH 7 (Marschner, 1984). There was an intermittent period of rainfall during the drainage period at 52–55 d after waterlogging (which is why redox decreased over this period). The horizontal black bar indicates waterlogging duration; the standard error of the mean of individual values was always less than ±20 mV.

Element toxicities in shoots of plants grown in soils during waterlogging

Soils from target environments exposed to waterlogging in Western Australia and potting mix were used to evaluate changes in microelement concentrations of plant shoots during waterlogging using three varieties (Fig. 5). All varieties in waterlogged Katanning soil had shoot concentrations of Mn, Fe and Al that were above the critical concentrations for toxicity (see Discussion); plants in waterlogged Esperance soil were high in Fe and Al; plants in waterlogged South Stirlings (Warburton) soil were high in Fe only; and one variety in waterlogged potting mix was high in Al (Fig. 5). In comparison, all varieties grown in these three soils under drained conditions (Camm, Fig. 5C) were usually below the critical levels for toxicity; however, in potting mix two varieties were marginally high in Mn and Al (Camm and Cascades, Fig. 5A and C).

Element toxicities were also pronounced in shoots of wheat exposed to waterlogging for only 12 d in two Indian soils at pH 8·2 and 9·4 (Fig. 6). There were usually only small increases in shoot Mn during waterlogging relative to drained plants (44 ± 28 mg kg⁻¹ shoot dry weight for waterlogged plants averaged across both soils), and these were usually below critical concentrations for toxicity (100 mg kg⁻¹ shoot dry weight). However, shoot Fe was high in both drained and waterlogged plants, and it increased to 440 ± 146 and 612 ± 237 mg kg⁻¹ (averaged across eight varieties) during waterlogging at pH 8·2 and 9·4, respectively. These shoot concentrations are 4–6 times higher than critical concentrations for toxicity. These same varieties showed significant increases in shoot Al and B concentrations during waterlogging in soil at pH 8·2, and these were 2- to 5-fold above critical concentrations for toxicity at 50 and 10 mg kg⁻¹, respectively (Fig. 6A and B). When the varieties were grown at pH 9·4, plants exposed to both drained and waterlogged conditions had similar high shoot concentrations of B and Al, and these were up to 5-fold above critical concentrations for toxicity (Fig. 6A and B). Concentrations of Na in the youngest three blades were also above critical concentrations for toxicity in waterlogged soil at pH 9·2, and also in drained soil at this pH (Fig. 6C).

Genetic diversity for tolerance to Al under aerated conditions

The genetic diversity of Australian and Indian wheat varieties for tolerance to high Al concentrations was evaluated using aerated solution culture. This showed a 3- to 4-fold genetic diversity for Al tolerance, with varieties like HD2009 and Ducula-4 having low Al tolerance, and varieties like Westonia and KRL 19 having high Al tolerance (Fig. 7). The correlation (r²) between root lengths in solutions without Al (–Al; control) and solutions at high Al was only 17 %.

Fig. 7.

Varietal tolerance to high Al concentrations based on mean maximum root length. Varieties were exposed to high Al concentrations (12 ppm) in aerated solution culture for 10 d (Materials and methods). Vertical lines indicate standard error of the mean.

DISCUSSION

Experiments presented here focus on waterlogging tolerance of wheat based on large germplasm sets (varieties and a doubled haploid population) evaluated in the field or in soils from waterlogged target environments that have been screened in the field in Australia and India. There are two key messages from this and earlier published research: (1) the waterlogging tolerance (ranking) and biomass production of wheat varieties under waterlogging is highly dependent on the environment and specifically the soil (see Introduction); and (2) waterlogging tolerance is not exclusively associated with differences in anaerobic conditions of soils, but rather is a product of anaerobic conditions and the traits associated with individual soils, such as elements and especially microelement toxicities (and possibility deficiencies) which are often exacerbated during waterlogging.

Data in Table 1 can be used to predict that element toxicities might occur in many of the soils used here based on critical concentrations for several elements and EC. These data suggest that Indian soils could be affected by high EC/salinity, and possibly by B, Mn and Fe toxicities, especially during waterlogging (Table 1). In comparison, the Australian soils could be affected by widespread Al and Fe toxicities, with some potential Mn toxicities (Table 1).

Minimum critical concentrations for toxicity based on soil analyses listed in Table 1 are at best only a general risk assessment for potential toxicities. For example, the concentration of total soluble Al may not be a suitable soil test because it does not consider the activity and form of the soluble Al which can be largely affected by soil conductivity (Carr et al., 1991; Carr and Ritchie, 1993; Barton and Carr, 1993). In drained alkaline sodic soils (pH ≥9·0) there are also large increases in shoot Al and reductions in plant growth when soils contain >0·8 mg kg⁻¹ of anionic Al [Al(OH)₄⁻; Ma et al., 2003; cf. 2–4 mg kg⁻¹ critical concentrations in Table 1]. Similarly for B, there are significant differences in critical concentrations for toxicity of soil soluble B for wheat that range from 27 to 53 mg kg⁻¹ in alkaline saline soils in Australia (Nuttall et al., 2006; cf. 3 mg kg⁻¹ critical concentrations in Table 1). There can also be large seasonal variations where concentrations of ions like Mn and Al may vary by 7- to 9-fold (for acid soils in Australia; Slattery and Ronnfeldt, 1992). Finally, any interpretation of data on soil analyses and potential critical concentrations would have to be considered in relation to soil pH (see discussions by Marschner, 1984; Atwell et al., 1999).

Data based on plant analyses demonstrate that waterlogging in Australian and Indian soils results in significant increases in shoot concentrations of Mn, Fe, Na, B and Al above critical concentrations for toxicity. Values for Al toxicity in shoots cited here of 50 mg kg⁻¹ shoot dry weight are 2–3 times higher than those where reduced growth occurred for wheat in sodic alkaline soils used by Ma et al. (2003) (14–21 mg kg⁻¹) because there were unknown effects of bicarbonate for the soils used here.

Increases in shoot Al, Mn, Fe and Na above toxicity concentrations have been shown for wheat waterlogged in another acidic Australian soil from Holly Siding (Khabaz-Saberi et al., 2006; waterlogged for 49 d). In this soil, the correlation between shoot Al concentration and biomass at the end of waterlogging was r = –0·92. In the experiments presented here and experiments of Khabaz-Saberi et al. (2006) there were little or no increases in shoot B during waterlogging.

The Australian soils used here were specifically selected to have little or no salinity (see EC and Na values, Table 1) thus minimizing the interactions of waterlogging and salinity known to occur for wheat (Introduction). This is different from plants waterlogged in the Indian soils where leaves were above critical concentrations for Na toxicities (Results and Fig. 6C).

In summary, there are ten observations supporting the adverse effects of element toxicities in wheat in waterlogged soils in Australia and India.

1. Redox potential measurements which fall rapidly in these soils (Ponnamperuma, 1972, 1984; Kirk et al., 2003; N. P. S. Yaduvanshi et al., unpubl. res.; Fig. 4) can be used to predict potential toxicities due to Mn, Fe and S (Marschner, 1984).

2. Soil analyses (Table 1) and particularly DTPA extracts for Mn and Fe in waterlogged soils confirm these increased levels (Ponnamperuma, 1972, 1984; Sparrow and Uren, 1987; Kirk et al., 2003; N. P. S. Yaduvanshi et al., unpubl. res.; Table 1).

3. Plant analyses (ICP) are above critical concentrations for Mn, Fe, Na, B and Al based on published data and data presented here (see Table 5).

4. Published information from toxicities for other crops grown in the region.

5. Recent findings of high Al concentrations in alkaline soils at these pH values (e.g. Ma et al., 2003; Rengasamy, 2004).

6. High microelement (Al) tolerance occurring in wheat from India and Australia selected for waterlogging prone areas but not for microelement tolerance (Fig. 7).

7. Elimination or minimizing adverse effects of waterlogging by elimination of soil microelements using potting mix (Khabaz-Saberi et al., 2006; Table 4).

8. Microelement indicator varieties and near-isogenic lines for B and Al which show differential growth in some of these soils when waterlogged but not when drained (Setter et al., 2006).

9. The observation that waterlogging effects can be minimized by changing pH to neutrality; that some wheat varieties also grow better in waterlogged alkaline soil than in drained soil (pers. obs. in both India and Australia) is consistent with waterlogging changing soil pH towards neutrality (Ponnamperuma, 1972, 1984) and thus minimizing microelement concentrations or effects (Marschner, 1984; Atwell et al., 1999).

10. Findings of high concentrations of microelements (Fe and Mn) on roots of wheat grown in waterlogged soils (Musgrave and Ding, 1998; see Results) or by simple visual examination of the orange iron oxide precipitates on wheat roots (Musgrave and Ding, 1998) or rice roots of plants grown in these same soils during waterlogging (pers. obs.).

Table 5.

Microelement toxicities known to occur for temperate cereals during waterlogging

Element	Crop	WL duration	Conditions	Reference
Mn	Wheat	Season	Natural WL gradients; loams; Australia	Sparrow and Uren, 1987
		2–6 d*	Sodic soil; India	Sharma and Swarup, 1988
		≥4 d*	Sodic soils; India	Swarup and Sharma, 1993
		≥28 d	River silt; USA	Ding and Musgrave, 1995
		17 d	Sand; USA	Huang et al., 1995
		Up to 100 d	Mn + Fe + P in roots; clay soil; USA	Musgrave and Ding, 1998
		49 d	Acidic soil; Australia	Khabaz-Saberi et al., 2006
		49 d	Acidic soil (Katanning)	Fig. 5A
	Barley	1–3 d*	Sodic soils; India	Bandyopadhyay and Sen, 1992
	Oat	21 d	Neutral soils; Denmark	Bjerre and Schierup, 1985
Fe	Wheat	2–6 d*	Sodic soil; India	Sharma and Swarup, 1988
		≥4 d*	Sodic soils; India	Swarup and Sharma, 1993
		≥28 d	River silt; USA	Ding and Musgrave, 1995
		17 d	Sand; USA	Huang et al., 1995
		49 d	Acidic soil; Australia	Khabaz-Saberi et al., 2006
		49 d	Acidic soils; Australia	Fig. 5B
	Barley	1–3 d*	Sodic soils; India	Bandyopadhyay and Sen, 1992
	Oat	21 d	Neutral soils; Denmark	Bjerre and Schierup, 1985
Na	Wheat	7 d	Sand; Australia	Barrett-Lennard, 1986
		2–6 d*	Sodic soil; India	Sharma and Swarup, 1988
		12 d	Sodic soil (pH 9·2); India	Fig. 6C
B	Wheat	12 d	Sodic soil (pH 8·2); India	Fig. 6B (pH 8·2)
Al	Wheat	49 d	Acidic soil; Australia	Khabaz-Saberi et al., 2006
		49 d	Two acidic soils; Australia	Fig. 5C
		12 d	Sodic soils; India	Fig. 6A

References are only cited where (a) plants are grown in waterlogged relative to drained soil; and (b) there is evidence that element concentrations in shoots grown in waterlogged soils are above the critical concentrations for toxicity during waterlogging as per Figs 4 and 5 (see Discussion). WL, Waterlogging.

* Waterlogging may have extended 2–3 weeks beyond this due to slow drainage of these soils (N. P. S. Yaduvanshi et al., unpubl. res.)

A working hypothesis to explain why such diverse microelements affect plant growth was originally presented by Setter et al. (2004) (Fig. 8). Results presented here indicate that tolerance to waterlogging in soil is a product of tolerance to anaerobiosis and high microelement concentrations (toxicities and perhaps deficiencies) which are associated with different soils. Waterlogging may affect microelement concentrations in soils directly by changes in soil redox potentials; or it may affect plant roots through either changes in energy (ATP) supply or membrane integrity. Similar effects on roots have been used to explain salinity and waterlogging interactions (Barrett-Lennard, 2003a). The influence of element toxicities specifically affecting growth is supported by diverse effects of waterlogging on growth in different soils or potting mix (Table 4), even though there are similar effects on redox potentials and the timing when soils would be anoxic after waterlogging, i.e. <350 mV at pH 7 as per Marschner (1984) (see also Fig. 4).

Fig. 8.

Effects of waterlogging on element toxicities due to (a) direct effects of low redox potentials on changes in concentrations (affecting concentrations of Fe²⁺, Mn²⁺ and S^2– ); (b) indirect effects of anaerobiosis on root energy supply and/or membrane integrity affecting the ability to exclude or compartmentalize ions such as Al, B and Na.

Earlier experiments conducted by E. G. Barrett-Lennard et al. (unpubl. res.; cf. Barrett-Lennard et al., 1999) demonstrate an interaction of hypoxia and salinity for wheat grown at up to 120 mm NaCl. It is not easy to confirm such interactions based on short-term treatment periods of ≤30 d, therefore data presented here are based on 33 d of hypoxia and salinity treatments followed by 13 d of recovery under aerated conditions (Fig. 9). At 0 mm NaCl, the shoot dry weight under hypoxia is approx. 50 % of aerated plants over the recovery period; however, at intermediate NaCl concentrations (≥20 mm NaCl) hypoxia reduced shoot dry weights to 0 % of aerated plants, i.e. plants died (as supported by visual observations). The dashed line in Fig. 9 shows the predicted shoot dry weights assuming there is no NaCl–hypoxia interaction based on the relative reductions of aerated plants at high NaCl. This line is substantially higher than what is measured under hypoxia with 20–120 mm NaCl, therefore confirming a NaCl–hypoxia interaction. Ethanol-insoluble dry weight is used in these analyses so as to eliminate any possible effect of the accumulation of salts affecting the dry-weight measurements.

Fig. 9.

Shoot ethanol-insoluble dry weight (EIDW) of wheat grown at a range of NaCl concentrations (0–120 mm; 0–12 dS m⁻¹) under either aerated or hypoxia conditions (nitrogen bubbled). Plants were grown under aerated or hypoxia conditions for 33 d and then allowed a further 13 d recovery (as per Barrett-Lennard et al., 1999); data plotted are for EIDW increases over the recovery period. The dotted line depicts the predicted results for hypoxia treatment (a) assuming that hypoxia reduces EIDW as per values at 0 mM NaCl, and (b) based on the relative reductions of the aerated treatment at 0–120 mm NaCl, i.e. assuming there is no NaCl –hypoxia interaction. The standard error of the mean is ≤10 % of individual values.

It is this interaction of salinity/NaCl and waterlogging/hypoxia that is proposed here to be similar to a wide range of other element toxicities which can be exacerbated by waterlogging or hypoxia. Key elements identified in this research relevant to toxicities during waterlogging in Australian and Indian soils include Al, Fe, Mn, Na and B. This is consistent with a large amount of published literature on waterlogging tolerance of cereals in diverse soils (Table 5). B and Na toxicities have often been highlighted as constraints in sodic soils (Sakal and Singh, 1995; Bell, 2005), and B toxicity has long been known to affect production of other crops in India due to high concentrations of B in groundwater used for irrigation (Sakal and Singh, 1995). However this is the first time that data have indicated that B toxicity affects wheat production, and that this is exacerbated in some soils during waterlogging. Whether Al toxicity also affects crops in these areas is less certain but the possibility is supported by recent findings of Al toxicity in highly alkaline sodic soils in Australia (Ma et al., 2003; Rengasamy, 2004; authors' unpubl. res.).

Other mechanisms of tolerance during waterlogging

It is highly likely that there are exceptions to the interactions of elements and anaerobiosis proposed here for wheat during waterlogging in some environments or soils. This may occur especially in neutral soils or in soils where there are little or no background element constraints. For example, preliminary ICP analyses data from India using plants grown in one waterlogged neutral soil indicate little support for element toxicities during waterlogging (pers. obs.). In this case one would expect a strong correlation between aerenchyma development and waterlogging tolerance; however, this did not occur for plants in waterlogged neutral soil (Rane et al., 2007). Some support for exceptions for element toxicities always being involved during waterlogging come from observations of positive relationships between tolerance to stagnant agar and waterlogging in a soil from Katanning, Western Australia (r² = 0·89 for 11 varieties; D. E. Goggin et al., unpubl. res.).

Phytotoxins are known to affect a large number of native plants grown in long-term waterlogged environments, since phytotoxins accumulate at redox potentials lower than where anoxia occurs (about 350 mV) or where Mn²⁺ and Fe²⁺ accumulate (about 350–450 and 150 mV, respectively, all at pH 7; Marschner, 1984). It would be easy to characterize wheat genotypes for tolerance to phytotoxins based on the pioneering work in this area in the UK (Armstrong et al., 1996a, b; Armstrong and Armstrong, 1999, 2001, 2005). As far as is known there is no published information on tolerance of a range of wheat varieties to phytotoxins. This is an important area for further work since phytotoxins are likely to be important in a wide range of soils during long-term waterlogging, i.e. either with or without element toxicity problems.

Priorities for future research

Research on element toxicities during waterlogging is the greatest priority for further work on waterlogging tolerance. First, even though the potential for element toxicities is clearly shown from soil data presented here, there is still no conclusive evidence that specific elements, or combinations of element toxicities, affect the tolerance and ranking of varieties exposed to waterlogging in particular soils. This remains a priority for further research. Secondly, the soil critical concentrations for element toxicities that have been used here extensively to highlight the potential impact on crop growth during waterlogging are based on plant growth under aerated conditions, e.g. Reuter et al. (1997). A top priority therefore is to determine whether soil critical concentrations for element toxicities are substantially lower for conditions where plants are waterlogged; this is the subject of continuing research. There appears to be no published information in support of this for any element, although this would seem likely for at least Na (see Fig. 9).

Thirdly, the finding of high Al in shoots of wheat grown in drained and waterlogged sodic soils in India (Fig. 6A) requires confirmation and is particularly important, because it specifically relates to mechanisms of improvement for waterlogging tolerance. Such effects are consistent with research highlighting high concentrations of total soluble Al [as Al(OH)₄⁻ at high pH] and Al toxicity in sodic alkaline soils in Australia (Ma et al., 2003; Rengasamy, 2004). This may also be an explanation for why reported waterlogging tolerant varieties like Ducula-4 do not express tolerance in some soils of India or Australia (Introduction), since Ducula-4 is highly intolerant to Al even under aerated conditions (Fig. 7). If confirmed, the findings of Al toxicity as a major constraint to crop production in sodic soils of India could be substantial since (a) between 5·7 Mha (in the Indo-Gangetic Plain only) to 6·9 Mha of soils are affected by sodic alkaline conditions in India (total areas across India; from tables 5 and 6, respectively, of Bhargava and Kumar, 2004), and these soils are highly prone to waterlogging; and (b) germplasm improvement for Al tolerance is based on simple screening protocols (Raman et al., 2005) with high heritability (Lafever and Campbell, 1978). Preliminary experiments support that indicator varieties for microelement toxicities could also be a cost effective alternative to extensive ICP analyses for plants and soils in these areas (see discussion by Setter et al., 2006).

Fourthly, this project has focused on element toxicities during waterlogging, yet there are also well-known examples of microelement deficiencies affecting plant, animal and human health in these Indian soils (Samra, 2006). There are also examples of microelement deficiencies including B associated with intolerance to waterlogging (see review by Jackson and Drew, 1984). Through a simple, comprehensive evaluation of the crop mineral status using ICP, both toxicities and deficiencies can be readily identified and then prioritized for either germplasm improvement or soil management approaches. Such approaches are especially important for identifying potential mechanisms of improvement for waterlogging tolerance.

Finally, with the confirmation of microelement toxicities during waterlogging based on plant tissue analyses in crops like wheat (Ding and Musgrave, 1995; Huang et al., 1995; Khabaz-Saberi et al., 2006) this work supports a priority to pyramid genes for multiple microelement tolerance. Such information will also affect the approaches for molecular improvements for waterlogging tolerance to extend from a focus on anaerobiosis (Dennis et al., 2000) to now include element toxicities. Recent approaches to develop molecular markers for waterlogging tolerance confirm that different QTLs exist for a doubled haploid wheat population exposed to waterlogging in different soils (Cakir et al., 2005). Nevertheless, in wheat (a) there is clear genetic diversity for waterlogging tolerance, e.g. Fig. 2; (b) waterlogging tolerance is often highly heritable in the field or different soils (Cai et al., 1996; Boru et al., 2001; Collaku and Harrison, 2005; Singh et al., 2006); and (c) there are often single or dominant additive genes involved in tolerance to element toxicities, e.g. for B (Moody et al., 1993; Campbell et al., 1994) and for Al (reviewed by Raman et al., 2005). Transcription factors might also be involved in any response of crops to waterlogging where multiple or complex mechanisms of tolerance are required.

Recent research on the Sub1A gene for submergence tolerance in rice supports the possible involvement of transcription factors (Fukao et al., 2006; Xu et al., 2006) in tolerance of plants to excess water. Submergence tolerance is a highly complex trait involving tolerance to multiple gas diffusion limitations (including anaerobiosis), limited light, and factors interacting with growth and phenology (Setter et al., 1997); however, a single transcription factor (ethylene-response factor-like gene; Sub1A) controls this complex response in indica rice varieties (Fukao et al., 2006; Xu et al., 2006; Perata and Voesneck, 2007). Whether such factors may be involved in tolerance of temperate cereals and other crops exposed to waterlogged soils remains the subject of continued research.