Increasing soil salinization decreases plant growth and impacts crop productivity world-wide. Of particular concern is the salinization of irrigated land, as this provides nearly half of the world’s food (Hopmans et al., 2021). An increase in the global demand for food for a growing population means that crops will increasingly be grown on salt-affected soils. Saline soils are defined as having an electrical conductivity (EC) greater than 4 dS m^-1 (40 mM NaCl), which can lead to a reduction in crop growth of 15 – 20% (Munns and Tester, 2008). However, saline soils can be much higher than this; salt distribution in soils is never uniform across an area and will vary with depth making it difficult to precisely calculate crop yield losses (Munns et al., 2020a). Glycophytes are salt sensitive species, tolerating only low concentrations of salt and include our most staple food crops (ie. wheat, maize, rice and barley). Halophytes are highly salt tolerant, growing in NaCl concentrations up to 500 mM. Compared with glycophytic plants they can accumulate high concentrations of salt (Na⁺ and Cl^-) in their leaf tissues to balance water relations (Colmer et al., 2006). Salt tolerance varies in glycophytic species, with barley and wheat moderately salt-tolerant compared to both Arabidopsis and rice (Munns and Tester, 2008). Salt tolerance can also vary within a species.

In terms of global food production, the cereal crops wheat (Triticum sp. L.), maize (Zea mays L.), rice (Oryza sativa L.) and barley (Hordeum vulgare L.) are the four most important crops. Wheat, grown in temperate regions, is the largest food crop in the world, providing approximately 20% of human calories and protein (Tadesse et al., 2019). Maize is grown predominantly in the Americas for grain, fodder, and as raw materials for industry (production of biofuels) (Shiferaw et al., 2011). Rice is a dietary staple for more than half of the world’s population. The Global Rice Science Partnership is predicting a 26% increase in rice production will be required in the next 25 years to meet global food demands (Zeigler and Dobermann, 2011). Together rice, wheat, and maize account for more than 50% human calorific consumption (FAO, 2022) and this is closer to 60% for developed countries (Awika, 2011). Barley is the fourth most important crop and is primarily grown for malting and livestock feed. It is the most salt-tolerant species out of the cereal crops, followed by wheat (moderately-tolerant), maize (moderately-sensitive) and rice (sensitive) (Munns and Tester, 2008), and is therefore an excellent model species for studying salt-tolerance mechanisms in crop plants.

Climate change is likely to accelerate soil salinization, because of the increased crop water requirements by elevated temperatures, through sea level rise, and the reduction in freshwater available for irrigation (Hopmans et al., 2021). The global impact of the changing climate is seen in interactions and feedbacks between climate, land degradation and food security. A significant increase in productivity of all cereal crops on salinized land is needed.

Plant roots have a pivotal role in providing anchorage, as well as the uptake of water and nutrients from the soil, for the maintenance of plant growth. The roots adapt to changes in the environment and soil conditions through many mechanisms, including the activity of membrane transporters that control the uptake of ions and water from the soil, their transport through the roots, and the loading into the xylem (Schroeder et al., 2013). In response to saline soils, the roots must exclude nearly all the salt from the soil solution (97 - 98%) whilst continuing to take up water by transpiration (Munns et al., 2020b). Plants exhibit root plasticity and can modulate their root system architecture (RSA) in response to their environment including hostile soils ie. drought, salt, flooding and extreme temperatures (Karlova et al., 2021). RSA also responds to soil nutrient deficiencies including nitrogen (N) and phosphorous (P) (Lynch, 2019). The steep, cheap and deep root ideotype, consists of physiological, anatomical and architecture traits for improved N acquisition in maize (Lynch, 2019).

Root traits have been studied less than shoot traits, mainly due to the inherent difficulties of imaging and quantifying root growth in soils. Modern root phenotyping technology has led to significant advancements in selecting root traits for improved crop productivity with promising root traits selected for abiotic stress tolerance incorporated into new germplasm for breeders. These include narrow xylem vessels to conserve water, shallow roots for surface P uptake, and longer roots for deep water capture (Tracy et al., 2020). Soil-surface roots (SOR) may also be beneficial for waterlogging tolerance in upland species such as rice to avoid hypoxic conditions (Hanzawa et al., 2013). Complex traits such as salinity tolerance are multigenic, therefore it is difficult to identify candidate genes for incorporation into breeding programs. Exploiting natural genetic diversity within a crop species by screening varieties, wild genotypes and landraces will be key for breeding more salt tolerant crops in the future (Shelden and Roessner, 2013). Genes for exclusion of Na⁺ from the xylem stream flowing to leaves in saline soil have been transferred into roots of modern wheat (Munns et al., 2012), but no specific trait for RSA that enhance salt-tolerance have been reported.

In this review we highlight the mechanisms that control crop root growth and development in response to salinity. We explore how this knowledge can be used to improve root system architecture for crop growth and yield in saline conditions. With significant gains in high resolution spatial profiling technologies and phenotyping, these tools can be used to understand salt-tolerance mechanisms from the individual cell to the whole plant system. The discovery of traits for root system adaptation to salinity stress will provide an under-utilised trait for breeders to develop more salt-tolerant crops.

Cereal crops are monocots, and their root structure comprises both embryonic (primary and seminal) and post-embryonic (brace, nodal and lateral) roots ( Figure 1 ). Monocots and dicots are characterised by different root structures, with dicots having an embryonic primary root (PR), lateral roots (LR) and adventitious roots (AR) (Smith and De Smet, 2012). Monocots have a fibrous root system and shoot borne roots that overtake the embryonic PR and LR and are mostly responsible for water and nutrient uptake. Barley, wheat, and maize have a variable number of seminal roots (including the primary root), with up to seven in some barley cultivars (Bengough et al., 2004), however this number can vary within species, and can also be affected by abiotic stresses. In contrast, the rice root system is comprised of a single embryonic seminal root, crown, and lateral roots. In maize and sorghum, below-borne roots are often referred to as crown roots and above ground brace (or prop) roots (Hochholdinger et al., 2004; Maqbool et al., 2022). In other cereals such as barley and wheat the shoot-borne roots are referred to as nodal roots. In cereals, the seminal root system is only active early in the growing period (up to the 4-leaf stage in barley and wheat).

Root hairs are specialised tubular structures that arise from epidermal cells. They are crucial for nutrient and water capture as they extend into the soil and increase the root surface area. The growth and development of root hairs are modulated by environmental signals, ensuring optimal acquisition of nutrients. The availability of less mobile nutrients such as inorganic P (Pi), iron (Fe), zinc (Zn) and magnesium (Mg) can all affect root hair morphogenesis (Lynch et al., 2014; Salazar-Henao et al., 2016). Root hairs are also important for the development and stabilization of the rhizosphere (McCully, 1999), and they can enhance penetration of high-strength soils (Haling et al., 2013). Root hair growth and development can also be affected by abiotic stresses, with saline soils shown to decrease both root hair length and density (Wang et al., 2008; Robin et al., 2016). In barley, root hairs also have been shown to contribute to drought tolerance and improved yield stability (Marin et al., 2020). Root hairs form part of the rhizosheath, along with soil particles bound to root hairs, mucilage and the root microbiome (McCully, 1999). Salt stress can influence the constitution of soil bacterial communities (Rath et al., 2019). Understanding the impact of saline soils on root-associated microbial communities will be important in generating crops with root systems adapted to salinity (Qin et al., 2016).

Salinity inhibits root system growth less than shoot growth, and the root:shoot biomass ratio increases in most species including sorghum (Weimberg et al., 1984), wheat (Husain et al., 2004), and rice (Lin and Kao, 1996). At low salinity, root system growth may not decrease at all while shoot growth declines as in barley and sorghum, or it may even increase as in bermudagrass and Cynodon turf grass (reviewed by Munns and Termaat (1986)). This may be due to stimulation of the rate of individual root elongation, as reported for cotton (Gossypium hirsutum L.) plants in saline solution by Kurth et al. (1986). In both barley and wheat, genetic variation has been observed in the primary root response to salinity (Rahnama et al., 2011; Shelden et al., 2013).

Elongation rate of the primary root was reduced by salinity up to 150 mM NaCl in Arabidopsis (Wu et al., 1996), barley (Shelden et al., 2013), wheat (Rahnama et al., 2011) and maize (Bernstein and Kafkafi, 2002). However, in cotton seedlings, the growth of the primary root was enhanced by moderate salinities (25 to 100 mM NaCl), and the cortical cells were longer than in control plants (Zhong and Lauchli, 1993). This increase occurred only when the Ca²⁺ concentration in the solution was supplied at 10 mM Ca²⁺ (Zhong and Lauchli, 1993). The rate of cell production in the presence of 10 mM Ca²⁺ was not affected by salinities up to 150 mM NaCl (Kurth et al., 1986). These findings reinforce the importance of having sufficient Ca(NO₃)₂ in the medium to prevent the Ca²⁺ activity falling in the presence of NaCl. The addition of NaCl or any electrolyte decreases the activity (the concentration of the free Ca²⁺ ion). For instance, addition of 150 mM NaCl to a solution containing 2 mM Ca(NO₃)₂ (i.e. half-strength Hoagland solution) decreases the activity of free Ca²⁺ to 0.64 mM (Husain et al., 2004). Raising the Ca²⁺ activity from 0.64 to 3.15 mM by increasing the Ca(NO₃)₂ from 2 to 10 mM in the presence of 150 mM NaCl resulted in a 50% increase in root system biomass in wheat (Husain et al., 2004). Calcium is an essential nutrient. It is a constituent of cell walls and affects many membrane functions including membrane permeability to cations such as Na⁺ (White and Broadley, 2003). The Ca²⁺ activity should be close to 1 mM (Genc et al., 2010). The Ca²⁺ concentration in irrigation water and groundwater of saline soils is higher than this (Rengasamy, 2010); salt treatments in laboratory experiments should replicate the natural environment when possible.

How plant roots “sense” Na⁺ in the soil and initiate the salt-specific response is still unknown. In animals, distinct protein receptors have been shown to be involved in sensing salt, with several ion channels found to act as salt sensing taste receptors (Chatzigeorgiou et al., 2013; Zhang et al., 2013). In roots, the sensor may be external to the epidermal cells, and detected by the proteins or lipids on the cell membrane. Alternatively, it could be within the cell and respond to small increases in cytosolic Na⁺. It is also assumed that plants must have mechanisms to detect both osmotic (water deficit) and ionic (salt-specific) stress. However, limited progress has been made because of the common usage of ‘salt shock’, a transfer of plants growing in non-saline hydroponic solutions to very high concentrations of NaCl or mannitol. This immediately causes a very large but transient rise in the concentration of cytosolic Ca²⁺. In Arabidopsis, OSCA1 is a hyperosmolality-gated calcium-permeable channel that results in increased cytosolic Ca²⁺ (Yuan et al., 2014). Structural analysis of OSCA family members have shown these belong to a mechanosensitive ion channel. In rice, cryo-electron microscopy (Cryo-EM) structure and function have shown how the rice homolog OSCA1 can mediate hyperosmolality sensing and transport pathway gating (Maity et al., 2019). Homologs of OSCA1 have been identified in the genome of rice (Li et al., 2015), maize (Cao et al., 2020), wheat (Tong et al., 2022) and soybean (Liu et al., 2022a). However, the role of OSCA1 as an ‘osmosensor’ is not yet proven as most experiments were done with very high concentrations of NaCl or mannitol which would cause plasmolysis, that is, extensive shrinkage of the protoplast with stretching of the plasma membrane between the sites where it remains tethered to the cell walls (Munns, 2002).

Salt stress sensors include both osmotic stress and Na⁺ sensors and may include pectin in cell walls, ion transporters, mechanosensory proteins, purine receptors, annexins and voltage-gated channels (Shabala et al., 2015; Wang et al., 2022). In the model plant Arabidopsis, AtIPUT1 that encodes an inositol phosphorylceramide glucuronosyltransferase, is required for the synthesis of glycosyl inositol phosphorylceramide (GIPC) sphingolipids, that is proposed as a potential salt sensor (Jiang et al., 2019). GIPCs are a class of glycosylated sphingolipids that are highly enriched in the outer leaflet of the plasma membrane (Mortimer and Scheller, 2020). GIPCs have been detected in barley using lipodomics (Yu et al., 2018), and in cotton, GIPC profiles change in response to salinity (Liu et al., 2022c). GIPCs are anchored in the plasma membrane and are activated by Ca²⁺ influx, leading to the up-regulation of Na⁺/H⁺ antiporters. Overexpression of IbIPUT1 in sweet potato (Ipomoea batatas (L.) Lam), resulted in a decrease in the accumulation of Na⁺ in root cells in response to salt stress (Liu et al., 2022b).

The highly conserved salt overly sensitive (SOS) pathway, first discovered in Arabidopsis is undoubtedly critical for salt tolerance (Shi et al., 2000). The SOS pathway consists of SOS1, a plasma membrane localised Na⁺/H⁺ antiporter ( Figure 3 ); SOS2, a CBL-interacting protein kinase (CIPK) family kinase, and SOS3, a calcineurinB-like (CBL) family Ca²⁺-binding protein. The SOS pathway is well characterised in Arabidopsis and is thought to confer salt tolerance by regulating net Na⁺ uptake at the soil-root interface. Although it has been extensively studied in Arabidopsis there is still relatively little known in crop plants. HvSOS1 expression has been shown in the root stele and epidermis of barley roots (Zhu et al., 2016). In maize, a 4-bp frame-shifting deletion was identified in ZmSOS1 impairing the function of SOS1, leading to a salt hypersensitive phenotype. A decrease in transcript level of ZmCBL8 (a CBL/SOS3 component of maize SOS pathway) also led to increased shoot Na⁺ content under salt conditions, indicating that the SOS pathway is crucial for salt tolerance (Zhou et al., 2022).

Root system architecture (RSA) refers to the spatial arrangement of the whole root system in the soil medium, modulated by root angle, type and rate of growth (Koevoets et al., 2016). The root surface area is also dependent on the development of root hairs and root diameter. RSA is complex and is considerably diverse among species, genotypes, and in space and time (Lynch, 1995). RSA is a plastic trait and responsible for establishing plant anchorage, optimising water and nutrient acquisition for plant growth, and avoiding unfavourable soil conditions (Smith and De Smet, 2012). Although there is significant research conducted on RSA in Arabidopsis, there is still limited research on the effect of salinity and other abiotic stresses on RSA in crop plants. One of the main limitations for understanding RSA is the difficulty in visualising root systems in soil.

Gravitropism is central to root growth as it is the gravity stimulus that drives the direction of root growth, with gravity sensing known to occur in the columella cells in the root tip. Salinity can interfere with the gravitropic response. Negative halotropism is the response of roots to grow away from salt. This could be a trait that is specific for salt-sensitive species, since in the halophytes (salt-tolerant species) Bassia indica and Limonium bicolor, the opposite phenotype – where roots grow towards salt, termed “positive halotropism”, has been observed (Shelef et al., 2010; Leng et al., 2019). Whether halotropism exists in crops, rather than root growth being inhibited by decreased water potential of the soil (due to drought or salinity), is unknown, and the costs to the plant modifying root architecture are unexplored (Munns and Gilliham, 2015). For roots to grow away from salt at depth, the exposure to salt must challenge or modify the plants response to gravity, causing the root growth direction to change. In the salt-sensitive species, tomato and sorghum, root bending away from salt is a strategy to avoid stress (Galvan-Ampudia et al., 2013). Root bending was only seen at > 300 mM NaCl concentrations indicating that the salt concentration needs to be relatively high to induce a halotropic root response.

Much of the research conducted on halotropism and root system architecture traits have been in the model plant, Arabidopsis. In Arabidopsis, the negative halotropic response was shown to be salt-specific ie. the roots did not change direction due to an equivalent osmotic gradient (induced by mannitol or KCl). The molecular mechanism for halotropism is due to the redistribution of the hormone auxin through the regulation of the auxin protein PIN2 (Galvan-Ampudia et al., 2013). Halotropism is dependent on the asymmetric redistribution of the hormone auxin from the side of the root closest to the salt, to the side of the root furthest from the salt. The salt-specific response involves changes in expression and localisation of the two auxin transporters, PIN2 (influx) and AUX1 (efflux) in the root tip (van den Berg et al., 2016). Arabidopsis plants overexpressing SOS1, showed a reduced halotropic response and Arabidopsis sos1 and sos2 mutants showed halotropic responses at lower thresholds of NaCl (Galvan-Ampudia et al., 2013). Taken together, these observations provide evidence that high internal concentrations of cellular Na⁺ may trigger the directional root growth, however, this has been unexplored experimentally. Differences in Na⁺ sequestration in the roots of durum and bread wheat, have been linked with differences in salt tolerance, however it is unknown if internal salt concentrations correlate with the halotropic response (Wu et al., 2018a). Whether other crop species, such as barley and wheat, have similar halotropic responses and share the same regulatory mechanisms is unknown.

Soil salinity, in both natural and managed environments (rain-fed and irrigated), is highly heterogeneous, and understanding how plants respond to this spatial and temporal heterogeneity is important for sustainable agriculture in more marginal lands. Split-root studies have shown how plants compensate for water and or nutrient deficiency, by optimizing root-foraging in the most favourable parts of the soil. This also applies to a saline soil. A review of halophytes grown with roots split between low and high salinity media, versus uniformly saline media, showed that water uptake from the least saline part of the soil or solution was the key factor driving shoot growth. Plants with part of the root at low salinity (0–10 mM NaCl) had 3- to 10-fold higher shoot dry mass than plants with roots in uniformly saline (50–800 mM NaCl) media (Bazihizina et al., 2012).

Plants exclude 97-98% of the salt, as remarked in the Introduction, so unless the soil receives frequent rainfall, or irrigation, the salt will build up to concentrations that prevent uptake further uptake of water. This implies that roots in saline soils without irrigation must continually forage for low-salt patches while balancing the need for nutrients. The root system architecture must be flexible enough to continue in situ if there is a history or ‘memory’ of frequent rain or irrigation that dilutes the salt that builds up around the roots or continue the exploration for low salt patches. The latter could become very expensive in terms of carbon cost.

Apoplastic barriers such as suberin lamellae and casparian bands are thought to limit the movement of salt transversely across the root by mass flow. In barley, salt stress reduces the hydraulic conductivity in the system by reducing water flow along the transcellular path (Knipfer et al., 2021). Transcriptomic and metabolomic analyses showed over-representation of genes and metabolites involved in the phenylpropanoid biosynthesis pathway in barley roots (Ho et al., 2020). Genetic differences have been observed in the formation of apoplastic barriers in barley genotypes (Ho et al., 2020) and rice (Krishnamurthy et al., 2011). Salt stress resulted in an increase in the production and abundance of apoplastic barriers in the endodermis of the Australian domesticated cultivar Clipper (Ho et al., 2020). In contrast, high levels of suberin synthesis-related gene expression and localization of suberin in cortical cells was observed in the elongation zone of the landrace Sahara, most likely restricting apoplastic transport of Na⁺ ( Figure 3 ). At 100 mM NaCl, Clipper accumulated higher concentrations of Na⁺ in the root maturation zone compared to the landrace Sahara (Shelden et al., 2013; Shelden et al., 2020). In both genotypes, K⁺ strongly correlated with Na⁺ concentration. A similar result was found in a transcript and proteomics analysis by Yamaguchi and Sharp (2010) with maize roots grown at low water potentials in vermiculite (low water content), suggesting that the formation of apoplastic barriers due to suberin and lignin deposition in cell walls of the stele was a response to low water potential. The deposition of suberised barriers in rice roots correlated with a sharp reduction in the uptake of Na⁺ in the shoot (Krishnamurthy et al., 2011).

An increase in cortical cell diameter was shown to occur in response to drought in wheat and is believed to lead to a reduction in energy usage (Colombi et al., 2019). Decreasing the number of cortical cell files by one or two can also reduce the energy costs associated with Na⁺ exclusion (Foster and Miklavcic, 2019; Munns et al., 2020a). Enlargement of root cortical cells in response to salt stress has been reported in barley (Shelden et al., 2016), maize (Li et al., 2014), cotton (Kurth et al., 1986) and halophytes (Hajibagheri et al., 1985) and may be due to increased vacuolation. Vacuolation is an adaptive mechanism to facilitate osmotic adjustment by the compartmentalisation of Na⁺ and Cl^-.

Plasticity in the development of root epidermal cells and differentiation of root hairs is thought to be a response to sensing environmental signals and adjusting to the stress response. In Arabidopsis, root hair density and length decreased in a dose dependant manner in primary roots (Wang et al., 2008). In both barley and wheat, root hair density and length were significantly reduced by salinity (Shabala et al., 2003; Robin et al., 2016), however osmotic stress has been shown to increase root hair density (Robin et al., 2021). Munns et al. (2020a) speculated that root hair cells may sequester salt into vesicles that are moved to the cell surface and extruded into the soil like what occurs in epidermal bladder cells of halophytes.

Membrane transporters located in the root tissues are responsible for the uptake of both Na⁺ and Cl^- ions from the soil, and for the uptake of water by aquaporins (Chaumont and Tyerman, 2014). Understanding the spatial and temporal expression of salt and water transporters is important to gain a complete understanding of salinity tolerance and root adaptation. It is also necessary to gain insights into how we can engineer crops to have improved root systems adapted to specific environments. At any point in time there are changes in the cell-specific spatial response to stress. Our ability to measure these changes are now enhanced by having access to high resolution ‘omic technologies for spatial profiling of metabolites, gene transcripts, proteins and ions (Shelden and Roessner, 2013).

The spatial and temporal variation in the expression of different transporters involved in water and nutrient transport in crop root tissues has been previously reported by Arsova et al. (2020), however our knowledge in this area is still limited, especially in crop plants. Salt overly sensitive (SOS1), a Na⁺/H⁺ antiporter, has previously been shown to be expressed in the epidermal cells surrounding the root apex in Arabidopsis (Shi et al., 2002) and barley (Zhu et al., 2016). Root epidermal cells contain Na⁺ transporters such as HvHKT1;1 in barley (Han et al., 2018). The distribution of membrane transporters associated with salt and water transport varies with both root age and space however mapping of transporters is biased towards Arabidopsis and young roots as information is lacking from other species and older roots (Arsova et al., 2020).

There are several genes encoding membrane transporters that have proven potential for increasing salt tolerance in crops. These include Nax2 from wheat (Munns et al., 2012), SKC1 (OsHKT1;5) from rice (Ismail and Horie, 2017) and the HAK family in maize seedlings (Zhang et al., 2019). Na⁺ exclusion genes, Nax1 (TmHKT1;4-A2) and Nax2 (TmHKT1;5-A) from the diploid wheat ancestor Triticum monococcum were introgressed into durum wheat Tamaroi and have both been shown to remove Na⁺ from the xylem in roots (Munns et al., 2012). Nax2 is expressed in the stele, specifically in the plasma membrane of the parenchyma cells lining the xylem vessels (Munns et al., 2012) ( Figure 3 ). Nax1 can also remove Na⁺ from the xylem flowing through leaf sheaths (James et al., 2011). Nax2 in Tamaroi resulted in higher grain yields in response to both salinity (Munns et al., 2012) and sodicity (Genc et al., 2016). TaHKT1;5-D Kna1 locus is localised on the plasma membrane of the root stele of bread wheat (Triticum aestivum L.) and retrieves Na⁺ from the xylem (Byrt et al., 2014).

Whilst there is a significant amount of research on root adaptation to salinity in the model plant Arabidopsis, we now need to focus our efforts on expanding our knowledge in crop species. Many of our major crop species are monocotyledons (ie. wheat, barley, rice, and maize) and have different root types and anatomical features compared to the dicot Arabidopsis. It is therefore likely they will have different mechanisms to adapt to saline soils. Understanding the root response to salt stress is complicated as it is dependent on species, genotype, plant age, and the intensity and duration of the stress. Many root studies have also been carried out in the laboratory in pots filled with soil or on media with artificial substrates, with very limited field studies. The advent of new high-throughput phenotyping platforms and technologies will help us to gain a better understanding of the diversity of root responses to saline soils across species. In the future, it will also be imperative to assess root traits in the field, and capture root responses to heterogenous soils.

Exploiting root traits with specific ideotypes in crop breeding has been shown to be successful for nitrogen acquisition and water capture in maize (Lynch, 2019). Salinity is a multigenic trait and identifying a specific root ideotype for salinity tolerance will be more challenging. We propose a salt-tolerant root ideotype for designing a root system adapted to saline soils. The salt-tolerant root ideotype would include root plasticity to allow roots to avoid highly saline soils whilst maximising water and nutrient uptake, transport mechanisms to exclude Na⁺ from the root and root anatomical traits to restrict Na⁺ movement into the xylem and conserve energy ( Figure 3 ). The identification of specific root traits adapted to saline soils and with current advances in crop genetics, these traits can be incorporated into crop breeding programs, leading to improved yields on saline soils.

MS and RM both contributed to writing the article. All authors contributed to the article and approved the submitted version.