Roots of Lucerne Seedlings are More Resilient to a Water Deficit than Leaves or Stems

Drought is one of the most harmful environmental stresses affecting the physiological, biochemical processes and growth of plants. Lucerne or alfalfa (Medicago sativa L.), one of the most popular pasture species in arid and semi-arid regions, plays a critical role in sustaining agricultural systems in many areas of the world. In order to evaluate the effect of water shortage on water status, biomass distribution and proline content, the relative water content (RWC), biomass and proline concentration in the leaves, stems and roots of lucerne seedlings under three different water regimes were studied in pots under a rainout shelter. The results showed that after water was withheld, the RWC of the different organs decreased significantly; at the same soil water content, the leaf RWC was higher than that of the stem and root. The biomass of the leaves, stems and roots were all reduced by water stress, while the root–shoot ratio increased indicating that the roots were less affected than the leaves and stems. Proline concentration increased with decreasing soil water content with the leaf proline concentration increasing more than that of stems and roots. These results indicate that roots of lucerne seedlings show greater resilience to water deficits than shoots.


Introduction
Lucerne is one of the most popular pasture species in arid and semi-arid regions and plays a critical role in sustaining agricultural systems by diversifying crop rotations, providing highly-nutritious livestock feed, enhancing soil quality through biological nitrogen fixation [1][2][3], and benefiting soil and water conservation [4].Precipitation in arid and semi-arid areas limits the agricultural production of crops and pastures; maximizing the efficiency of precipitation use in the production of biomass is therefore important in the arid and semi-arid regions of northwest China [5].Drought also affects the growth and development of lucerne [6], limiting its distribution.In China, the planted area of lucerne is the largest among forage species and has been grown for at least 150 years on the Loess Plateau [7].Xinjiang Daye is the cultivar of dryland lucerne most widely planted in the arid and semi-arid areas of northwest China due to its high productivity, its adaptation to drought and its ability to survive the cold winters.Autumn and spring are often the seasons when lucerne is sown, but as the winter temperatures are often much lower than 0 • C on the Loess Plateau [8], autumn sowing usually results in unsuccessful survival during winter.Thus on the Loess Plateau, lucerne is best sown in April, but precipitation is very limited in spring, so if the rainfall of the previous year is low, lack of water after the seed germination can affect the growth and survival of the lucerne seedlings.Plants adapt to drought by a series of phenological, physiological, morphological and biochemical processes [9].The ability of plants to maintain favorable water status and efficiently use available resources is critical for the growth and survival in water-limited environments.The leaf relative water content (RWC) and leaf water potential are the indicators of the degree of water stress and cellular hydration of the plant tissues [10].While the RWC and water potential of leaves are often measured to determine the plant's water status, the water status of roots and stems are only infrequently measured.As water moves through the plant along a potential gradient from the roots to the leaves, the water potential and RWC of the leaves are expected to be lower than that of the stems and that of the stems lower than that of the roots during daylight hours, but the RWC of roots and stems of lucerne subjected to water deficits have not been measured as far as is known.
As lucerne is a forage legume, the biomass accumulation of the shoots is an important measure of productivity.Under drought stress, differential effects have been observed among lucerne genotypes for biomass accumulation and harvestable yield [11,12]. However at the seedling stage, root growth may be more important than aboveground biomass accumulation for the establishment of the lucerne pasture, particularly in water-limited environments.While drought stress reduces both root and shoot growth, root growth in plants is often less affected than the shoots leading to an increase in the root-shoot biomass ratio (R:S) [13,14], but the degree of partitioning of biomass between different organs of lucerne is largely unknown.
Silva [15] reported that plants accumulate both organic and inorganic solutes in the cytosol to raise the osmotic pressure and thereby maintain both turgor and the driving gradient for water uptake.Among these solutes, proline is the one of the most widely studied [16,17]; with water shortage, proline is largely accumulated in the cytosol to balance the accumulation of solutes in the vacuole, a process known as osmotic adjustment [18].However, the role of proline in the establishment of lucerne under water-limited conditions is unknown.
Although there is some research on the morphological, physiological and ecological characteristics of lucerne under soil water stress [19][20][21], studies of the relationships among relative water content, biomass and proline accumulation in roots, stems and leaves of lucerne are very limited.The main objective of this study was to evaluate the effect of water stress at the seedling stage on the RWC, biomass and proline accumulation.The hypotheses of this research were as follows: (1) the growth of lucerne will be inhibited by water stress, but the different tissues will be affected to different degrees; (2) water stress will affect the biomass distribution so as to increase the R:S ratio; (3) proline concentration of the tissues will act as a marker for the degree of water deficit in the tissues.

Materials
Medicago sativa L. cv.XinjiangDaye was selected as the experimental material; seeds were kindly provided by Gansu Academy of Agricultural Sciences.Twelve plastic pots, diameter 350 mm and height 400 mm with holes at the bottom to allow for drainage, were each filled with 9.5 kg of air-dry sandy-loam top soil collected from the campus of Gansu Agricultural University to which 0.5 kg of livestock manure was added.The soil had a pot capacity (PC) of 25.7%, determined by wetting up the soil until free draining from the pot and then allowed to drain for 24 h before weighing.Before sowing, 1 L nutrient solution (NH 4 NO 3 : 5.42 g/L, KH 2 PO 4 : 2.33 g/L) was added to each pot and 20 seeds per pot were sown on 1 April 2017.The 12 pots were placed under a rainout shelter (3 m long × 5 m wide × 2.7 m high that was closed only during rain events) in the experimental field of the College of Forestry at Gansu Agricultural University (36 • 09 N, 103 • 70 S, altitude 1620 m), Lanzhou, Gansu Province, China.Except during the infrequent rain events, the plants were grown under natural light, temperature and humidity conditions.

Methodology
A preliminary experiment showed that if the lucerne seeds had not emerged within seven days from sowing, they were unlikely to emerge thereafter.On 8th April (the average temperature and precipitation in April in Lanzhou is 13 • C and 30 mm), the plants were thinned to three uniform plants per pot and three water treatments, each with four replicates, were imposed and arranged in a randomized complete block design: (i) TA, full water supply (the minimum soil water content was 85% PC); (ii) TB, moderate water stress (the minimum soil water content was 65% PC); and (iii) TC, severe water stress (the minimum soil water content was 45% PC).The soil water content (SWC) reached the predetermined levels within 24 h.Using the gravimetric method to control the SWC, every pot was weighed at 17.30 h Beijing Standard Time (BST) each day and watered to maintain the SWC at the three predetermined levels: 85%, 65% and 45% PC.After 10 days (18 April), the pots were destroyed between 10.00 and 12.00 h BST, the plants were cut off at soil level, soil quickly washed from the roots, the plants rinsed with water, and then wiped dry with paper towel.Two plants per pot were used to measure RWC and biomass, while the remaining plant was used to measure the proline concentration.
After separation into roots, stems and leaves, the fresh weight (FW) of each tissue was determined within 300 s, floated in distilled water for 6 h to gain turgidity, then the turgid weight (TW) was determined.The turgid leaves, stems and roots were then oven-dried at 80 • C for 24 h.The dry weight (DW) of the leaves, stems and roots were measured to determine their biomass and to calculate the RWC of each component: Total biomass = leaf biomass + stem biomass + root biomass The third plant was also quickly separated into leaves, stems and roots to measure the proline concentration of each tissue using the method of Bates et al. [22].First, 0.5 g of fresh leaf, stem and root were homogenized in 5 mL 3% sulfosalicylic acid (1.25 g ninhydrin, 30 mL of glacial acetic acid, 20 mL of H 3 PO 4 ), and incubated for 600 s at 100 • C.After cooling, the mixture was filtered to a test tube and the filtrate used to determine the proline concentration.The filtrate (2 mL) was mixed with equal volumes of glacial acetic acid and ninhydrin reagent and heated for 0.5 h in a boiling water bath.Four mL of toluene was added after cooling, shaken for 30 s before taking the upper layer of fluid into a 10 mL centrifuge tube and centrifuging at 14,000 g for 300 s.The absorbance value of the upper proline/toluene solution was measured spectrophotometrically at 520 nm using free toluene as a blank control.The proline concentration was determined using a standard curve, and expressed as µg g −1 fresh weight.

Statistical Analysis
The general analysis of variance (ANOVA) was used to test for significance of all measured variables using SPSS 12.0, Comparisons of means among organs and between treatments were carried out by the Tukey's test at P < 0.05 (Table 1).All correlations were done by Pearson analysis.

Effect of Water Treatments on the Relative Water Content of Leaves, Stems and Roots
Ten days after the imposition of the three water treatments, the mean root, stem and leaf RWC decreased significantly with decreasing water supply from 85% at full water supply (85% PC) to 82% at moderate water stress (65% PC) to 79% at severe water stress (45% PC) (Figure 1a).The mean RWC of the root (79%) was lower than that in the stem (82%) and lower than that in the leaves (85%) (Figure 1b).There was no interaction between tissues and water treatments.

Effect of Water Treatments on the Relative Water Content of Leaves, Stems and Roots
Ten days after the imposition of the three water treatments, the mean root, stem and leaf RWC decreased significantly with decreasing water supply from 85% at full water supply (85% PC) to 82% at moderate water stress (65% PC) to 79% at severe water stress (45% PC) (Figure 1a).The mean RWC of the root (79%) was lower than that in the stem (82%) and lower than that in the leaves (85%) (Figure 1b).There was no interaction between tissues and water treatments.

Effect of Water Treatments on the Biomass of Leaves, Stems and Roots
Total biomass decreased from 0.44 g plant −1 under full water supply to 0.33 g plant −1 in the moderate stress treatment to 0.11 g plant −1 in the severe stress treatment (Table 2).The total biomass of the seedlings subjected to the severe water stress treatment (45% PC) was significantly lower than that of the seedlings with the full water supply (85% PC) and moderate stress treatment (65% PC) that did not differ significantly between each other (Table 2).The root, stem and leaf biomass all decreased with decreasing soil water content, but there was a strong interaction between the level of water stress and the effect on biomass accumulation in the different tissues (Figure 2).The differences in biomass among different tissues were smaller with the decreasing SWC (Figure 2a) and the differences in leaf and stem biomass were significant among the three water treatments (Figure 2b), while the root biomass was not significantly different between full water supply (85% PC) and the moderate stress treatment (65% PC).After 10 days of growth the leaf biomass of the seedlings was There was no significant interaction between tissues and water treatments.

Effect of Water Treatments on the Biomass of Leaves, Stems and Roots
Total biomass decreased from 0.44 g plant −1 under full water supply to 0.33 g plant −1 in the moderate stress treatment to 0.11 g plant −1 in the severe stress treatment (Table 2).The total biomass of the seedlings subjected to the severe water stress treatment (45% PC) was significantly lower than that of the seedlings with the full water supply (85% PC) and moderate stress treatment (65% PC) that did not differ significantly between each other (Table 2).The root, stem and leaf biomass all decreased with decreasing soil water content, but there was a strong interaction between the level of water stress and the effect on biomass accumulation in the different tissues (Figure 2).The differences in biomass among different tissues were smaller with the decreasing SWC (Figure 2a) and the differences in leaf and stem biomass were significant among the three water treatments (Figure 2b), while the root biomass was not significantly different between full water supply (85% PC) and the moderate stress treatment (65% PC).After 10 days of growth the leaf biomass of the seedlings was 42% of the total biomass with the full water supply (85% PC), decreasing to 39% with a moderate stress treatment (65% PC) and 37% when subjected to the severe stress treatment (45% PC).By contrast, the root biomass was 23% of the total biomass when maintained at 85% PC, increasing to 27% when maintained at 65% PC and 30% when maintained at 45% PC, resulting in the R:S ratio increasing from 0.30 to 0.38 to 0.43 from the fully watered to moderate strass to severe stress treatments (Table 2), indicating that the growth and biomass allocation/accumulation of the roots was less affected than that of the leaves by the water deficit.
Table 2. Effect of three watering treatments (TA, well-watered (85% pot capacity (PC); TB, moderate water stress (65% PC); and TC, severe water stress (45% PC)) on the total biomass, the leaf, stem and root biomass as a percentage of the total biomass, and root-shoot ratio of lucerne seedlings.The values are means ± one standard error of the mean (n = 3).

Treatment
Total Biomass (g plant −1   Root to Total Biomass (%)

Effect of Water Treatments on the Proline Concentration of the Leaves, Stems and Roots
The mean concentration of proline increased as the SWC decreased in all three tissues (Figure 3a), while the proline concentration was higher in leaves than in the stems with the lowest concentration in the roots (Figure 3b).The leaf proline concentration increased significantly from the full water treatment (2.04 μg g −1 FW) to the moderate stress treatment (3.31 μg g −1 FW) to the severe stress treatment (4.28 μg g −1 FW), more than double that of the fully-watered treatment.This contrasts with the proline concentration in the roots that increased from 1.64 μg g −1 FW in the well-watered lucerne seedlings to 1.94 μg g −1 FW in the moderate stress treatment, but this was not significantly different from the fully watered treatment, at 2.52 μg g −1 FW, a 54% increase over the fully-watered treatment in the severe stress treatment (Figure 3b).The leaves therefore accumulated more proline than the roots in response to a water deficit.

Effect of Water Treatments on the Proline Concentration of the Leaves, Stems and Roots
The mean concentration of proline increased as the SWC decreased in all three tissues (Figure 3a), while the proline concentration was higher in leaves than in the stems with the lowest concentration in the roots (Figure 3b).The leaf proline concentration increased significantly from the full water treatment (2.04 µg g −1 FW) to the moderate stress treatment (3.31 µg g −1 FW) to the severe stress treatment (4.28 µg g −1 FW), more than double that of the fully-watered treatment.This contrasts with the proline concentration in the roots that increased from 1.64 µg g −1 FW in the well-watered lucerne seedlings to 1.94 µg g −1 FW in the moderate stress treatment, but this was not significantly different from the fully watered treatment, at 2.52 µg g −1 FW, a 54% increase over the fully-watered treatment in the severe stress treatment (Figure 3b).The leaves therefore accumulated more proline than the roots in response to a water deficit.3a), while the proline concentration was higher in leaves than in the stems with the lowest concentration in the roots (Figure 3b).The leaf proline concentration increased significantly from the full water treatment (2.04 μg g −1 FW) to the moderate stress treatment (3.31 μg g −1 FW) to the severe stress treatment (4.28 μg g −1 FW), more than double that of the fully-watered treatment.This contrasts with the proline concentration in the roots that increased from 1.64 μg g −1 FW in the well-watered lucerne seedlings to 1.94 μg g −1 FW in the moderate stress treatment, but this was not significantly different from the fully watered treatment, at 2.52 μg g −1 FW, a 54% increase over the fully-watered treatment in the severe stress treatment (Figure 3b).The leaves therefore accumulated more proline than the roots in response to a water deficit.

Correlation among Soil Water Content, Relative Water Content, Biomass and Proline Concentration
With a decrease in average SWC, root, stem and leaf RWC all decreased linearly (Figure 4a), with the root RWC decreasing more (0.21% per 1% decrease in SWC) than that of leaves (0.17% per 1% decrease in SWC) and the stem RWC decreasing the least (0.09% per 1% decrease in SWC) when grown at lower SWC (Table 3).Similarly, when grown at lower SWC, root, stem and leaf biomass all decreased linearly with average SWC (Figure 4b), with the leaf biomass decreasing at 0.0036 g per 1% decrease in SWC from an initial value of 0.18 g in the fully-watered treatment, followed by stem biomass decreasing at 0.0029 g per 1% decrease in SWC from an initial value of 0.15 g in the fully-watered treatment, and the root biomass decreasing at 0.0017 g per 1% decrease in SWC from an initial value of 0.10 g in the fully-watered treatment (Table 3).In contrast to the decrease of RWC and biomass at lower average SWC, proline concentration was negatively correlated with SWC (P < 0.05), that is, proline concentration increased as the lucerne seedlings were grown in soil maintained at lower values of SWC (Figure 4c).Although the proline concentration in the well-watered treatment (85% PC) was similar in all the plant tissues at 1.64-2.0µg g −1 , the increase in the proline concentration at the lower average SWC was greatest in the leaves (0.06 µg g −1 % −1 decrease in SWC), intermediate in the stems (0.04 µg g −1 % −1 decrease in SWC) and least in the roots (0.03 µg g −1 % −1 decrease in SWC) (Table 3).
The root, stem and leaf biomass each decreased linearly with proline concentration (Figure 4d) so that there were tissue-specific relationships between proline concentration and biomass accumulation with water availability (Table 3).
The root, stem and leaf biomass each decreased linearly with proline concentration (Figure 4d) so that there were tissue-specific relationships between proline concentration and biomass accumulation with water availability (Table 3).3.  3.
Table 3.The linear regressions and correlation coefficients between the average soil water content (% pot capacity (PC)) and relative water content (RWC), biomass and proline concentration, and the relationship between biomass and proline concentration in the leaves, stems and roots of lucerne seedlings.

Discussion
Exposure to three different average soil water contents for 10 days shortly after emergence resulted in lucerne seedlings having higher values of RWC in the leaves than stems and roots when adequately watered (maintained at 85% PC), but lower values of RWC in the leaves and stems in the severe water stress treatment (maintained at 45% PC) due to a smaller decrease in stem than leaf RWC with a decrease in SWC (Figure 4a).The unique relationship of the tissues between RWC and SWC was also true in biomass accumulation and proline accumulation over the 10 days in the different water treatments (Figure 4).As expected, biomass was reduced by moderate and severe stress in all the plant tissues with the decrease being greater in the leaves than roots, corresponding to a greater increase in proline in leaves than stems than roots.Thus the leaves, stems and roots responded in tissue-specific ways to the decrease in SWC, with leaf biomass and proline accumulation responding more and root biomass and proline accumulation less to a decrease in SWC (Figure 4).While the biomass accumulation over the 10 days differed considerably between leaves, stems and roots in the well-watered treatment, with leaves accumulating almost double the leaf than root biomass (Figure 2), the proline concentration was equally low in all tissues (Figure 3).Nevertheless, the rate of decrease in biomass with proline concentration was similar in all tissues (Figure 4) suggesting that the accumulation of proline with the development of water deficit was similar in all three tissues, so that the concentration was a function of the amount of biomass that was produced in the three tissues.While it is well recognized that root biomass accumulation is less sensitive to water deficits than leaf biomass accumulation, leading to an increase in the root-shoot ratio [23], the similarity in proline production/accumulation among the plant tissues with a soil water deficit has not been previously reported.
Haffani [24] found that shoots of Vetch species (Vicia narbonensis, V. sativa and V. villosa) were more sensitive than roots because their biomass reduction was greater.Similar results were reported by Porter [25] who reported that moderate and severe drought stress reduced the growth of shoots rather than roots, compared to the well-watered plants.Our results are in accordance with these studies.However, Talluto [26] had a contrasting conclusion that water deficit significantly increased root biomass.This can be considered as an adaptive strategy for plants to maintain root growth enabling the plant to improve water uptake [27].The decrease in biomass in all three tissues maintained at lower SWCs with a greater decrease in the leaves than roots is consistent with Hypothesis 1 that growth of lucerne will be inhibited by the reduction in SWC and that the R:S ratio will increase in the plants maintained at 65% PC and 45% PC compared with those maintained at 85% PC.
The differences in RWC between the watering treatments were relatively small, decreasing from 85.0% when the average SWC was maintained at 85% PC to 81.9% when the average SWC was maintained at 65% PC and 78.9% when the average SWC was maintained at 45% PC.This is a reflection of the watering regime of daily deficit irrigation [28].The plants were watered each day at 17:30 h and after 10 days of each treatment sampled for RWC at 10.00 to 12.00 h the following day after 11 h of darkness (sunrise and sunset in April in Lanzhou is 06.30h and 19.30 h, respectively).Nevertheless, the leaves stems and roots had significantly different values of RWC with leaves and stems having a higher RWC at 82-84% than the roots at 79% RWC.As leaves, stems and roots were taken as a whole for RWC measurement, the lower RWC in the roots is contrary to the normal gradient in a transpiring plant [29], probably reflecting the expected gradient in SWC down the pot from high SWC near the surface to low SWC at the base as a result of the daily watering from the top [30].However, the lower RWC in the roots may have been temporary and the RWC of the leaves in the afternoon may have been considerably lower than when measured in the morning.The decrease in the accumulation of leaf biomass and increase in leaf proline accumulation were both much larger in the leaves than stems than roots with decreasing water availability (average SWC), consistent with a greater mean water deficit over the 10 days in the leaves than stems and roots.The fact is that the values of RWC represent values at a particular time-of-day on the day of harvest while the biomass and proline are cumulative values from 10 days at three average SWCs.However, there was a relationship between the increase in proline concentration on a fresh weight basis with decrease in biomass on a dry weight basis in the three tissues (Figure 4d), but this was not a unique relationship among the tissues so that the small differences in the initial concentration of proline among the tissues in the fully-watered plants became much larger with the increase in the proline concentration as the increase in biomass was reduced in the plants maintained at low SWCs (Figure 4).
Taïbi [31] reported that free amino acids concentration could be used as a trait to identify the ability of Pinus halepensis in seedlings to cope with drought stress.The results of the present study showed that the concentration of proline in the roots, stems, and leaves increased under water stress Agronomy 2019, 9, 123 9 of 11 with both moderate and severe water restrictions.The accumulation of proline was greater in the leaves than the stems than the roots with decreasing average SWC.The role of proline has been widely discussed in several publications.Stewart and Hanson [16] argued that proline was a metabolite produced under stress conditions possibly because of inhibition of proline oxidation and impaired protein synthesis, but was not an adaptive response to water stress, while Aspinall and Paleg [32] reported that the accumulation of proline could be an indicator of drought resistance.Likewise, Sharma and Verslues [33] suggested that plants adapt to drought by the accumulation of non-toxic compounds, such as proline, that prevent cell damage at low cell water potentials and maintain the osmotic potential of the cytoplasm as solutes accumulate in the vacuole.As Silva et al. [15] pointed out, the osmotic pressure of the plant cell regulates many processes through the accumulation of proline inside the cell under drought conditions.As osmotic adjustment has been shown to be an adaptive mechanism that benefits yield in some crop species when exposed to drought [18,34], proline plays a role in drought adaptation and desiccation tolerance.Chiang and Dandekar [35] reported that the proline concentration was also dependent on plant age, leaf age, position, or leaf part (see also Verbruggen et al. [36]).While the measured RWC do not reflect the cumulative water deficit by the plants and plant tissues, the proline concentration does appear to reflect the cumulative stress over the 10 days with greater concentration in the leaves than stems than roots and a greater concentration in all three tissues with the lower volume of water applied over the 10 days.Thus the third hypothesis that the proline concentration of the tissues acts as a marker of the degree of water deficit of the tissue is upheld.

Conclusions
Our results have shown that the maintenance of the soil water content at average SWCs of 65% PC and 45% PC resulted in a decrease of RWC and biomass of the roots, stems, and leaves of lucerne seedlings compared with the fully-watered seedlings.Water deficiency also induced an increase in proline content in all these three tissues.Leaf biomass was affected the most under severe water stress followed by stem biomass, and the least affected was root biomass.Leaves were more sensitive to water stress than the roots and stems in that the biomass was reduced more in leaves and stems than roots, resulting in a higher R:S ratio when the average SWC was reduced.The proline concentration of the roots and stems did not increase as much as the leaves with the severe water treatment, again suggesting that the roots were more resilient than the leaves to a limited water supply.This study of physiological responses of lucerne to soil water stress during the seedling stage provides a foundation for understanding the basis of drought resistance in lucerne that can be used by agronomists and extension specialists to determine the timing, risks and opportunities for seedling establishment of lucerne in spring in semi-arid and arid environments.
same upper-case letter are not significantly different among different plant tissues (T) in the same water treatment.Means with the same lower-case letter are not significantly different among different water treatments (W) of the same plant tissue.T × W is the least significant difference (LSD) of the interaction between tissues and water treatments.

Figure 1 .
Figure 1.Effects of three water treatments (TA, well-watered (85% PC); TB, moderate water stress (65% PC); and TC, severe water stress (45% PC)) on the mean relative water content (RWC) of the leaves, stems and roots of lucerne seedlings (a), and the mean RWC of the tissues in the three water treatments (b).Means with the same upper-case letter are not significantly different among different plant tissues.Means with the same lower-case letter are not significantly different among different water treatments.There was no significant interaction between tissues and water treatments.

Figure 1 .
Figure 1.Effects of three water treatments (TA, well-watered (85% PC); TB, moderate water stress (65% PC); and TC, severe water stress (45% PC)) on the mean relative water content (RWC) of the leaves, stems and roots of lucerne seedlings (a), and the mean RWC of the tissues in the three water treatments (b).Means with the same upper-case letter are not significantly different among different plant tissues.Means with the same lower-case letter are not significantly different among different water treatments.There was no significant interaction between tissues and water treatments.

11 Table 2 .
± 4.2 b 33.0 ± 5.4 c 36.8 ± 6.3 b 0.43 ± 0.005 c Means with the same lower-case letter are not significantly different among different water treatments.Agronomy 2019, 9, x FOR PEER REVIEW 5 of Effect of three watering treatments (TA, well-watered (85% pot capacity (PC); TB, moderate water stress (65% PC); and TC, severe water stress (45% PC)) on the total biomass, the leaf, stem and root biomass as a percentage of the total biomass, and root-shoot ratio of lucerne seedlings.The values are means ± one standard error of the mean (n = 3).Means with the same lower-case letter are not significantly different among different water treatments.

Figure 2 .
Figure 2. Effect of three water treatments (TA, well-watered (85% PC); TB, moderate water stress (65% PC); and TC, severe water stress (45% PC)) on the root, stem and leaf biomass of lucerne seedlings (a), and the comparison of the effects of the three water treatments in the tissues (b).The bars are ± one standard error of the mean (n = 3).Means with different lower-case letters are significantly different at P = 0.05 according to Tukey's multiple range test.

Figure 2 .
Figure 2. Effect of three water treatments (TA, well-watered (85% PC); TB, moderate water stress (65% PC); and TC, severe water stress (45% PC)) on the root, stem and leaf biomass of lucerne seedlings (a), and the comparison of the effects of the three water treatments in the tissues (b).The bars are ± one standard error of the mean (n = 3).Means with different lower-case letters are significantly different at P = 0.05 according to Tukey's multiple range test.

Figure 3 .
Figure 3.Effect of three water treatments (TA, well-watered (85% PC); TB, moderate water stress (65% PC); and TC, severe water stress (45% PC)) on the proline concentration (fresh weight basis) of leaves, stems and roots of lucerne seedlings by treatment (a) and the comparison of the effects of the three water treatments in the tissues (b).The bars are ± one standard error of the mean (n = 3).Means with different lower-case letters are significantly different at P = 0.05 according to Tukey's multiple range test.

Figure 4 .
Figure 4. Relationship between average soil water content (% pot capacity (PC)) and (a) relative water content (RWC), (b) biomass, (c) proline concentration and (d) the relationship between biomass and proline concentration in the leaves, stems and roots of lucerne seedlings.The linear regressions and correlation coefficients are shown in Table3.

Figure 4 .
Figure 4. Relationship between average soil water content (% pot capacity (PC)) and (a) relative water content (RWC), (b) biomass, (c) proline concentration and (d) the relationship between biomass and proline concentration in the leaves, stems and roots of lucerne seedlings.The linear regressions and correlation coefficients are shown in Table3.

Table 1 .
Effect of three watering treatments (TA, well-watered (85% pot capacity (PC); TB, moderate water stress (65% PC); and TC, severe water stress (45% PC)) on the RWC of leaves, stems and roots.The values are means ± one standard error of the mean (n = 3).Differences were tested with Tukey's method.Means with the same upper-case letter are not significantly different among different plant tissues (T) in the same water treatment.Means with the same lower-case letter are not significantly different among different water treatments (W) of the same plant tissue.T × W is the least significant difference (LSD) of the interaction between tissues and water treatments.