Evaluation of morphological and phytochemical characteristics of Mesona chinensis populations in southern China

ABSTRACT Mesona chinensis Benth. is a herbal plant with both edible and medicinal values. Thirty-four populations of M. chinensis in southern China were examined to quantify their morphological and phytochemical characteristics. The results showed that the populations differed significantly (p < 0.01) in each of the morphological and phytochemical characteristics. The morphological characteristics showed wide variation among populations, with CV values ranging from 15.76% to 54.18%. The herbage yield was mainly positively correlated with a number of stem traits. The cluster analysis showed that the populations belong to two major groups. Both total phenolic (TPh) contents and total flavonoid (TF) contents respectively ranged from 5.49 to 18.44 and 5.06 to 11.47 mg⋅g−1 among populations. Antioxidant activity, evaluated by means of ABTS (2, 2-azino-bis-3-ethylbenzothiazoline-6-sulfonic acid) and DPPH (1,1-diphenyl-2-picrylhydrazyl) assays, ranged from 63.91 to 223.41 and 30.35 to 137.84 mmol Trolox equivalents⋅g−1, respectively, among populations. Significant positive correlations were found between antioxidant activity and TF contents, TPh contents, and total polysaccharide (TP) contents. Most of the 34 populations were rich in polysaccharides, while GD-6, GD-12, GD-3, GD-8, JX-6, and FJ-1 had the highest TP content in turn. Based on comprehensive grey correlation analysis, GD-1, GD-2, GD-8, FJ-6, FJ-3, FJ-2, JX-6, GD-7, JX-3, and JX-2 were the best performers for production and quality of M. chinensis in southern China.


Introduction
Mesona chinensis Benth. is a perennial herb with both edible and medicinal values (Wu & Raven, 1995). As a well-known traditional Chinese herbal medicine, it is commonly used to cure heatstroke and fever in traditional Chinese medicine (Liu & Fang, 1998), and treat diseases including diabetes, hypertension, and acute nephritis in modern medicine as well (Shen et al., 2000). The dry stems and leaves of M. chinensis are important raw materials for the production of herbal teas and black jelly (Lai & Chao, 2000). The herb is rich in polysaccharides (Liu & Chen, 2004). The polysaccharides obtained from the herbage of M. chinensis can improve the human immune system (Zhao et al., 2011), and it is also a good natural coagulator in the food preparation process, improving the mechanical properties of starch gums, and promoting the coagulation of starch and other substances (Yuris et al., 2018).
M. chinensis is widely cultivated in the southern provinces of China, such as Guangdong, Guangxi, Fujian, and Jiangxi. It is estimated that there are more than 10,000 hectares of M. chinensis cultivated in China (Shen et al., 2000). With the increasing demand for mesona herbal teas and black jelly, large-scale commercial production of this crop is being promoted. However, one of the problems facing the widespread domestication of this herb is the lack of superior cultivars for the different planting regions. In China, the landraces of M. chinensis are directly selected from different wild genotypes, while the large number of landraces have relatively low yield and quality. Such case increases the need for breeding new cultivars for the diverse South China environments.
The effective screening and multidirectional evaluation of M. chinensis genetic resources can help plant breeders to select valuable genotypes for further breeding projects (Nan et al., 2019). The morphological characterization of plant materials with the desired traits is essential to the effective use of crop germplasm (Santos et al., 2012), and can help breeders to select the most suitable genotypes for breeding programs (Goodarzi et al., 2018). A previous research showed that some morphological characteristics of M. chinensis were related to yield in different cultivation environments . However, there is no published report on the evaluation of either the morphological characteristics or yield among landraces or cultivated populations of M. chinensis. Phytochemical characterization is also regarded as an effective method for evaluating germplasm material of medicinal plants (Nsuala et al., 2017). The herbage of M. chinensis contains polysaccharides, polyphenols, and flavonoids (Lin et al., 2016). However, information on the accumulation of the main bioactive components in breeding lines of M. chinensis is still lacking. In addition, previous researchers reported that the aqueous extract of M. chinensis had antioxidant activity (Adisakwattana et al., 2014). However, the differences in the antioxidant effects among populations of M. chinensis have not been assessed or characterized. Given all the above, M. chinensis cultivars with excellent comprehensive traits are desirable for plantation and utilization.
The present study was to reveal the variation of 34 M. chinensis populations by examining their morphological and phytochemical characteristics, and identify desirable traits and superior populations for practical implementation and further breeding programs.

Sources and collection of planting materials
In the current study, 34 populations of M. chinensis were collected from four provinces of China during 2016 to 2017, which represent all the regions where M. chinensis is cultivated in China (Table 1). The 34 collection sites were selected because they were distributed across the length and breadth of the four provinces. All the individuals of each population were cultivated in the greenhouse of the South China Agricultural University (Guangzhou, Guangdong, China) to preserve the germplasm.

Experimental design and planting
To evaluate the variation in morphological and phytochemical characteristics available among and within the 34 M. chinensis populations, a common garden test was carried out in the experimental field located at the South China Agricultural University, Guangzhou, Guangdong, China (23°09′13.07″ N latitude, 113°16′4.51″ E longitude) during March to October, 2018. The average annual temperature and precipitation at this site are 22.7°C and 1732.2 mm, respectively. The field belongs to lateritic red soil and had been used previously for continuous growing of rice.
The common garden test was arranged in a completely randomized design, with three replications for each population. Due to its small seed size and low germination rate, the plants were propagated vegetatively by stem cuttings for the next generations. Each replicate plot contained 30 cuttings from one population and the planting area for each replicate of each population was 15 m 2 . The cuttings were rooted in the greenhouse and transplanted into the field by the time they had grown to a height of 30 cm. The cuttings were spaced 50 cm apart within rows and 60 cm apart between rows. Before planting seedlings, the field was fertilized with a compound fertilizer (containing 15% N, 15% P 2 O 5 , and 15% K 2 O) at a rate of 6 g m −2 of N. The experiment was conducted under irrigation, and weeds were controlled.

Morphological evaluation
Ten individual plants were randomly selected from each replicate plot for morphological analysis in October. In all, 16 morphological characteristics were evaluated as shown in Tables 2 and 3. Most of the characteristics were measured using a ruler. The herbage dry weight (per one plant) was determined on an electronic scale to a precision of 0.01 g. The branch number of the individual plants was counted. Leaf area (per one leaf) measurement was carried out by scanning the leaves, and then using the Image-Pro software (Media Cybernetics, USA) to measure the area of the leaf on the bitmaps. The leaf aspect ratio (each leaf) was determined as the leaf length divided by the leaf width. In addition, some qualitative characteristics were measured using a semi-quantitative scale (Table 3).

Phytochemical evaluation
Three individual plants were randomly selected from each replicate plot for chemical analysis in October. The aerial parts of the plant were dried in the oven, and were ground to fine powder (˂ 380 μm) by a Highspeed Mill (Dadeyaoji., China). They were stored in airtight packs for further chemical analysis.
The content of total polysaccharide (TP) was determined according to the method described by Zhang et al. (2019) with some modifications. Briefly, each sample (2.0 g) was homogenized in 60 mL of Na 2 CO 3 solution (1.5% w/v) at 95 °C for 4 h. The extract was filtered, diluted using water and precipitated overnight with absolute ethanol at 4°C, before the pellet was collected by centrifugation (2300 g × 15 min). The pellet was then dried, suspended in water and the polysaccharide content was determined using the phenol-sulfuric acid colorimetric method, using a Shimadzu UV1800 spectrophotometer (Shimadzu, Tokyo, Japan) at 490 nm. The content of TP was calculated as D-glucose equivalents and expressed as mg D-glucose equivalent·g −1 dry weight of plant material.
The content of total flavonoid (TF) was determined by the method of Sarker and Oba (2019), with some modifications. Briefly, 80 mL of 60% aqueous ethanol was added to 2.0 g of each ground dried plant sample. The suspension was then extracted by placing in a water bath at 70 °C for 2 h. Then, the extract was centrifuged (2300 g × 10 min), filtered, and diluted using 60% aqueous ethanol for assay of TF content. An aliquot of the diluted filtered extract (2.0 mL) was mixed with AlCl 3 (0.1 mL, 10%), potassium acetate (0.1 mL, 1 M) and distilled water (2.8 ml). The absorbance was read after 30 min at 415 nm. The content of TF was calculated as rutin equivalents and expressed as mg rutin equivalent·g −1 dry weight of plant material.
The content of total phenolic (TPh) was determined using the Folin-Ciocalteu colorimetric method described previously (Cai et al., 2004), with a minor modification. Briefly, 2.0 g of the ground sample was extracted with 70 mL 25% aqueous ethanol at 60 °C for 1 h in a water bath. Then the extract was centrifuged (2300 g × 10 min), filtered, and diluted using water. Each dilution (0.2 mL) of the filtered extract was mixed with 1 mL of 2 mol·L −1 Folin-Ciocalteu reagent, 4 mL of 20% aqueous sodium carbonate and 2.8 mL of distilled water. The absorbance of the resulting blue color was measured at 760 nm with a spectrophotometer after incubation for 45 min at room temperature. Gallic acid was used as the reference standard and the TPh content was expressed as mg gallic acid equivalent (GAE)·g −1 dry weight of plant material.
Antioxidant activity was determined according to the method described previously (Surveswaran et al., 2007), with some modifications. Briefly, for ethanolic extraction, 10 mL of 80% aqueous ethanol was added to 20.0 mg dried ground plant material in a conical flask, which was kept at room temperature overnight with occasional shaking. Then, the extract was filtered using a nylon membrane Millipore filter with 0.45 μm pore size at room temperature. The filtrate was stored at 4 °C before analysis. Using the method of Surveswaran et al. (2007) to prepare the ABTS (2, 2-azino-bis-3-ethylbenzothiazoline-6-sulfonic acid) radical (ABTS + ) solution (7 mM), 7.8 mL of ABTS + solution was mixed with 0.2 mL of sample extract and the reaction mixture was allowed to stand at 23 °C for 1 h. The absorbance was immediately measured at 734 nm. A Trolox standard curve was prepared and the results were expressed as Trolox equivalent antioxidant The DPPH (1,1-diphenyl-2-picrylhydrazyl) assay procedure was similar to the ABTS method described above. Using the method of Cai et al. (2003) to prepare the DPPH radical (DPPH + ) solution (0.06 mM), 7.8 mL of DPPH + solution (absorbance of 0.68 ± 0.005 at 515 nm) was added to 0.2 mL of ethanolic sample extract and the reaction mixture was carried out at 23 °C in the dark for 30 min. The absorbance was recorded at 515 nm. A Trolox standard curve was also prepared and the results were expressed as TEAC units (mmol TE·g −1 dry weight of plant sample).

Data analysis
The mean values of the morphological and phytochemical traits were calculated and used for statistical analysis. Mean, minimum value, maximum value, standard deviation (SD) and coefficient of variation [CV (%) = standard deviation/ mean ×100] were measured for each variable. SPSS software (IBM, Armonk, NY, USA) was used to analyze all the morphological and phytochemical traits, using one-way analysis of variance (ANOVA), Pearson's correlation analysis (to determine the relationship between variables) and principal component analysis (PCA). Frequencies and percentage distributions of the qualitative characteristics were  determined. The path coefficient analysis was used to evaluate the contribution of the morphological characteristics to herbage yield, using the method described by Dewey and Lu (1959). PCA was performed to investigate the relationships between the populations (Goodarzi et al., 2018). The grey relational analysis was used to evaluate the herbage yield and quality integratedly, using the method described by Nan et al. (2019) and Deng (2010). After calculating the correlation between the major system behaviors and the influencing factors, certain data are processed to form a quantitative description of the influence of the factors on the whole system (Nan et al., 2019). To better understand the patterns of variation among the populations, the Euclidean distance matrix was used to carry out cluster analysis, using the PAST statistics software (Paleontological Statistics) (Hammer et al., 2001). In addition, scatter plot analysis was created using PAST software on the basis of PC1 and PC2.

Morphological characteristics
Each of the quantitative morphological characteristics showed significant differences (p < 0.01) among populations. The estimates of morphological variation in individual populations are summarized in Table 2.
Of the stem characteristics, branch number varied from 28.5 to 98.5 with a mean of 60.3 among populations, whereas main stem length and branch length were within the ranges 48.5-112.1 cm (mean 74.1 cm) and 36.4-78.0 cm (mean 54.4 cm), respectively, and stem base width values ranged from 0.5 to 1.0 cm, with a mean of 0.7 cm. Of the leaf characteristics, leaf area varied from 4.5 to 9.4 cm 2 with a mean of 6.2 cm 2 , and leaf length and leaf width varied from 3.2 to 5.7 cm (mean 4.0 cm) and 1.6 to 3.0 cm (mean 2.3 cm), respectively. The qualitative morphological characteristics also showed variation in terms of frequency distribution in the studied populations (Table 3). Illustrations of growth habit, leaf shape, main stem color and stem pubescence of M. chinensis are shown in Figure S1.
The mean values and range of variation for each of those 16 characteristics are presented in Table 4. Most of the morphological traits had a relatively high CV value. The highest CV was observed for branch number (54.18%), followed by growth habit (53.06%) and herbage dry weight (51.95%), while the lowest CV was given for leaf length (15.76%). Out of the 16 morphological traits, 14 exhibited CV values greater than 20%. The main economic trait, herbage dry weight per plant, ranged from 5.70 to 385.50 g (mean value of 84.15 g) (Table 4).

Morphological diversity
A scatter plot was created with respect to PC1 and PC2, which accounted for 39.38% of the total variance ( Figure  S2). In the plot, clustering of populations together indicated their similarity. The populations GD-1, FJ-3, GX-1, GD-2, FJ-2 and FJ-6 were categorized into one group, with the remaining 28 populations being categorized into another group.
Hierarchical cluster analysis, based on the combination of both quantitative and qualitative morphological traits, grouped the 34 populations of M. chinensis into two main clusters in the dendrogram (Figure 1), reflecting the findings of the PCA plot ( Figure S2). The first group (I) was divided into two subgroups. Subgroup I-A contained one population (FJ-6), and subgroup I-B contained five populations (FJ-2, GD-2, FJ-3, GD-1 and GX-1). The second group (II) was divided into three subgroups. Subgroup II-A contained one population (GD-9), subgroup II-C contained two populations (JX-1 and FJ-5), and subgroup II-B contained the remaining 25 populations.

Herbage yield attributes
Path coefficient analysis was performed to assess the magnitude of the contributions of individual traits to herbage dry weight. The results of the analysis of direct and indirect contributions of various traits to herbage dry weight are shown in Table 5  (0.294). The highest negative direct effect on herbage dry weight was due to plant height (−0.287), followed by main stem internode length (−0.263). For indirect effects, branch length showed the greatest indirect effect on herbage dry weight via main stem length (0.306), followed by stem base width via branch number (0.289), stem base width via main stem length (0.189), stem base width via branch length (0.188), main stem internode length via main stem length (0.264), branch length (0.248), branch number (0.201) and stem base width (0.192). Therefore, branch number, main stem length, stem base width and branch length were shown to be the main direct and indirect contributors to herbage yield in M. chinensis.

Phytochemical characteristics and antioxidant activity
There were significant differences (p < 0.01) in TP, TPh, TF, ABTS, and DPPH among populations ( Table 6). All of the phytochemical characteristics and antioxidant activity exhibited variation among the 34 populations (Table 6). TP contents of the 34 populations varied from 52.48 to 101.44 mg.g −1 , with a mean of 77.81 mg.g −1 . The GD-6 population was found to have the highest TP content among the 34 populations, followed by GD-12 and GD-3, whereas FJ-2 population had the lowest. The content of TP in GD-6 was ~1.9 times that in FJ-2, showing the wide variation in TP content among M. chinensis populations.
TPh contents in M. chinensis populations ranged from 5.49 to 18.44 mg.g −1 plant dry weight, with a mean of 12.21 mg.g −1 . The JX-6 population had the highest TPh content, whereas the FJ-10 population had the lowest. The content of TPh in population JX-6 was ~3.4 times that of the FJ-10 population.
The range of TF contents in M. chinensis populations varied from 5.06 to 11.47 mg.g −1 plant dry weight. The mean content of TF among the 34 populations of M. chinensis was 8.02 mg.g −1 of plant dry weight. The GD-7 population had the highest content of TF, whereas the FJ-6 population had the lowest content of TF. The TF content of GD-7 was ~2.3 times that in FJ-6.
The total antioxidant activity of M. chinensis populations determined by the ABTS and DPPH assays are given in Table 6. The antioxidant activity determined by the ABTS assay showed wide variation, from 63.91 to 223.41 mmol TE·g −1 (mean 161.15 mmol TE·g −1 ). The GD-3 population had the highest antioxidant activity, as determined by the ABTS assay, whereas the FJ-6 population had the lowest. The antioxidant activity determined by the DPPH assay also showed wide variation, from 30.35 to 137.84 mmol TE·g −1 with a mean of 70.54 mmol TE·g −1 .
The JX-6 population had the highest antioxidant activity determined by the DPPH assay, whereas the FJ-6 population had the lowest.

Correlation between morphological, production, and phytochemical characteristics
Significant correlations (p < 0.01) were detected for a number of measured parameters (Table 7). Herbage yield (Herbage dry weight) showed particularly high positive correlations with a number of stem traits, such as stem base width, branch number, main stem length, branch length, main stem internode length, and branch internode length. No significant negative coefficient was observed between herbage dry weight and any other of the morphological traits studied.
The antioxidant activity determined by the DPPH assay showed a significant (p < 0.01) positive correlation with TF, TPh and TP, respectively. The antioxidant activity determined by the ABTS assay was significantly (p < 0.05) positively correlated with the TF.

Morphological characteristics
Each of the quantitative morphological characteristics of M. chinensis showed significant differences among populations (p < 0.01), and each of the morphological characteristics showed considerable variability among the 34 populations, with CV values ranging from 15.76% to 54.18%. In fact, high levels of variation have been widely found among clonally propagated species such as grape (Vitis vinifera) and seedless barberry (Berberis vulgaris var. asperma) (Dzhambazova et al., 2009;Goodarzi et al., 2018;Hvarleva et al., 2004). The phenotypic diversity in a plant germplasm could be due in part to variation in local conditions such as topography, soil conditions, weather, sunlight, rainfall and other outstanding features of the location (Goodarzi et al., 2018). In the current study, all the plants of the 34 different populations, collected from the major growing areas of China, were planted at the same time in the same field. The use of the common garden test minimized the impacts of environmental factors on the morphological characteristics of the various populations. It is reasonable to speculate, therefore, that genetic differences among the populations made the major contribution to the diversity of morphological characteristics among the 34 M. chinensis populations in the common garden test.
In selecting a potentially superior plant type, correlation studies provide reliable information about the nature and direction of the selection pressure (Srivastava et al., 2018). Our results revealed positive correlations between herbage dry weight and the stem traits. These findings are in accordance with an earlier investigation (Su et al., 2010). Furthermore, path coefficient analysis of morphological traits revealed that branch number, main stem length, stem base width, and branch length should be considered for direct selection in M. chinensis, as selection for these traits would be expected to bring an improvement in the herbage yield of the crop. These results showed that the herbage yield of M. chinensis was mainly determined by a number of stem traits. Thus, selecting for improved stem traits could be an effective strategy for improving the yield of the herbage crop.

Phytochemical characteristics
M. chinensis is an interesting multipurpose herb crop for tropical and subtropical climates. Regarding the chemical composition of the biomass, the scarcity of available information limits comparisons that would be very useful in elucidating the possible food potential of this herb (Sulas et al., 2016). In the current study, all of the phytochemical characters of M. chinensis showed significant differences (p < 0.01) among the 34 populations, which provided some valuable information for the improvement of the herb.
TPh and TF are important components in M. chinensis extracts. In previous studies, the contents of TPh in Data are presented as means of three replicates. ABTS: antioxidant activity determined by the ABTS assay (mmol TE.g −1 ), DPPH: antioxidant activity determined by the DPPH assay (mmol TE.g −1 ), TPh: total phenolics (mg.g −1 ), TF: total flavonoids (mg.g −1 ), TP: total polysaccharides (mg.g −1 ). Signifcance: **P < 0.01. Table 7. Correlation coefficients between morphological, yield and phytochemical characteristics of Mesona chinensis populations. M. chinensis were reported to be between 3.901 mg .g −1 (Kuang et al., 2012) and 185.4 mg.g −1 dried plant (Hung & Yen, 2002). The values were different from the mean value from the current study (12.21 mg.g −1 ). The varying contents of TPh may result from differences in cultivars, geographic origins, growing seasons, agricultural practices, and analytical methods (Kim et al., 2003). On the other hand, mean TF content in M. chinensis was reported to be 10.37 mg.g −1 dry plant weight in previous research (Kuang et al., 2012), which was similar to the results from the present study (mean of 8.02 mg.g −1 ). In our study, correlation analysis revealed that both the contents of TF and TPh were significantly correlated with antioxidant activities based on the ABTS and DPPH assays, implying that TF and TPh contribute to the antioxidant activities of this Chinese herb. In fact, many studies have documented the relationship between antioxidant activity and the content of both TPh and TF compounds (Sulas et al., 2016).
Mesona polysaccharides have been demonstrated to possess many medicinal properties, such as antioxidant activity (Lin et al., 2017), hypoglycemic activity (Huang et al., 2018) and immunoregulation (Lin et al., 2018). In particular, mesona polysaccharides have good gelling properties, and are unique in that they synergistically increase the viscosity and gel strength of certain nonwaxy starches, including those of wheat, maize, and rice, during the pasting process (Feng et al., 2014). These characteristics of mesona polysaccharides could expand their applications in starch-based production (Feng et al., 2013). TP content of M. chinensis was reported to be at the level of 8.1% to 9.6% under different cultivation conditions . That result was higher than the average value that achieved in the present study (mean of 7.78%). However, even the value we obtained was markedly higher than that reported from some other plants, such as TP contents of 1.07% to 4.13% in mulberry fruits (Khan et al., 2019) and 5.3% to  6.7% in Crocus sativus accessions from different origins . The results indicated that the herbage of M. chinensis was rich in polysaccharides and may be a good raw material for processing plant polysaccharides (Liu et al., 2019;Yuris et al., 2018).

Selection of superior germplasm
The herbage of M. chinensis was frequently used by Chinese people in the preparation of herbal tea or black jelly (Jiang et al., 2014). In China, the yearly consumption of herbal tea and black jelly made of mesona herbage is huge. At present, with the market for botanical beverages and functional foods gradually expanding, the need for mesona herbage has increased remarkably. Therefore, it is imperative to develop genetic materials with higher yield and quality for M. chinensis. Phenotypic characters are easily measured and frequently have high heritability, so that selection based on these traits would be appropriate for the rapid screening of plant breeding materials and improvement of performance (Yap & Harvey, 1972). In the current study, the results of grey relationship analysis showed that the 10 populations (GD-1, GD-2, GD-8, FJ-6, FJ-3, FJ-2, JX-6, GD-7, JX-3, and JX-2) had the highest production and quality performance in turn among the populations evaluated. Of them, some populations, such as GD-1, GD-2, GD-8, FJ-3, and FJ-2, had higher herbage yields with 25% above the average of all populations. Meanwhile, some populations, such as JX-2, JX-3, JX-6, GD-2, GD-7, and GD-8, had higher antioxidant activities determined by the ABTS assay or by the DPPH assay, with 20% above the average of all populations. Higher antioxidant activities in M. chinensis herbage contribute to higher product quality of herbal tea or black jelly. Although 'FJ-6ʹ seems to have normal herbage dry weight and relatively low phytochemical characteristics and antioxidant activity in all populations, it has an outstanding performance in other traits. Based on an integrated evaluation of grey relational analysis, the population is one of the best performers in all populations. Thus, we suggest that the 10 populations should be used as potential breeding materials for improving the yields and antioxidant activities of this Chinese herb.
To meet the needs of industrial extraction of mesona polysaccharides, the cultivars with high TP content is needed. In this study, TP content ranged from 5.25% to 10.14% in all the populations. Among the populations evaluated, GD-6, GD-12, GD-3, GD-8, JX-6, and FJ-1 had the highest TP content with 15% above the average of all populations. Thus, these populations are suggested as superior breeding materials to improve TP content of M. chinensis.

Conclusion
Among 34 populations of M. chinensis in southern China, the morphological and phytochemical characteristics showed significant differences (p < 0.01). The herbage yield was significantly correlated with stem traits. Both the contents of TF and TPh were significantly correlated with antioxidant activities. Most of the 34 populations were rich in polysaccharides, while GD-6, GD-12, GD-3, GD-8, JX-6, and FJ-1 had the highest TP content in the populations. Based on comprehensive grey correlation analysis, GD-1, GD-2, GD-8, FJ-6, FJ-3, FJ-2, JX-6, GD-7, JX-3, and JX-2 were the best performers for production and quality of M. chinensis in southern China. High morphological and phytochemical variations obtained in the present study suggest that there is genetic potential to develop M. chinensis with high yield and excellent herbage quality.