Study area

The research focused on the south-southeastern region of Mexico due to the presence of commercial cultivations of moringa and this characteristic favored seed availability. The seeds of the accessions studied are from the states of Veracruz, Oaxaca, Guerrero, Chiapas and Yucatán (Table 1, Figure 1). The research was carried out at the College of Postgraduates, Campus Veracruz, geographically located at 19° 16’ 32” north latitude, 96° 16’ 32” west longitude, at an altitude of 16 m (^{Valdés et al., 2014}). The climate is warm subhumid (AW₀), with an accumulated annual rainfall of 1 000 mm and an average annual temperature of 27 °C (^{Olguín, 1999}).

Figure 1 Geographical location of the collection points of M. oleifera in the South-Southeast of Mexico. 

Biological material

Moringa seeds were collected from different commercial cultivations in the South-Southeast of Mexico during the months from February to May 2018. Thirty healthy seeds were selected from each collection point and the variables of length, thickness and weight of the seed were measured. Twenty accessions were evaluated, which are shown in (Table 2).

Sowing and transplanting

The seeds were sown in black nursery bags of 27 x 27 cm. The substrate used was composed of soil, manure-vermicompost and sand (5:4:1). After 15 days of germination, the seedlings of greater vigor and of a similar size were selected and two months later, they were transplanted in the field. The type of soil in the field was clay-loam.

Treatments and experimental design

The experimental design was completely randomized blocks with five repetitions. The size of the area used was 900 m². The distance between individuals and furrows was 3 m. Every seven days, the variables of height, basal diameter (10 cm from the ground), number of branches and leaves were recorded. Two liters of water was applied daily to each plant through drip irrigation.

Morphological characterization

For morphological characterization, quantitative and qualitative descriptors were used for each organ of the plant. The quantitative ones recorded were: height, stem diameter, number of leaves, number of branches, number of flowers, leaf length, leaf width, petiole length, leaflet length, leaflet width and days to first flowering. For qualitative descriptors, those published by ^{Mgendi et al. (2011}); Zhigila et al. (2015); ^{Popoola et al. (2016}) were used. The qualitative descriptors were: leaf petiole color (1: light green, 2: green, 3: light violet, 4: medium violet, 5: intense violet), leaf shape (1: oval, 2: oblong, 3: oblong oval and 4: elliptical), leaf apex (1: obtuse and 2: acute), leaf pubescence (0: absent and 2: present), flower color (1: white, 2: white-cream, 3: white-pink, 4: white-cream-pink and 5: pinkish), purple spots on the flowers (0: absent and 1: present) and anther color (1: yellow and 2: orange).

Statistical analysis

The average values of the descriptors were evaluated by descriptive statistics. The correlation coefficient, principal component analyses and cluster analysis were performed through a hierarchical grouping with coefficient of variation in the unweighted pair group method (Euclidean distance). The two-dimensional scatter plot was performed through the percentage variation of the first two principal component analyses using the PAST program Version 3.0.

Germination

The highest percentage of germination was obtained in accession G3 (100%) and the lowest percentage was 13% for accessions C4, O1, G4, C1, C4 and Y1. At the time of transplantation, accession V3 measured 108 cm, being the greatest height. Significant statistical differences (p˂ 0.05) were observed in the height and diameter of the stem of the 20 accessions evaluated. Accession C3 had the largest diameter and the largest number of leaves corresponded to accession O4 (Table 3).

Great variation was found in the percentage of germination of the accessions sown. This variation is attributed to seed quality and storage time before sowing (^{Du Toit et al., 2017}). In addition to this, genetic potential and environmental factors such as temperature, precipitation and altitude influence the development of the seed before collection and determine its size and weight (^{Baiyeri et al., 2015}; ^{Ledea-Rodríguez et al., 2018}).

This process affects the viability of seeds of more than one year of storage in temperature ranges of 23 to 25 ºC. Therefore, it is recommended that the seeds be preserved in the pods (^{Fotouo et al., 2015}). The average number of germination days of the accessions sown was 11. The earliest gemination occurred at 8 days and the latest at 14 days. This interval is like the 6 and 13 days reported by ^{Popoola et al. (2016}); ^{Zaku et al. (2015}); while ^{Kumar et al. (2014)} mentioned that germination occurs between 10 and 12 days after sowing. ^{Ramos et al. (2010}) reported that germination begins from 8 days after sowing and at 25 days, the primary leaves appear.

Height

Accessions C2 and G2 were the largest and reached an average height of 273 and 271 cm, respectively. Accession Y4 had an average height of 164 cm, being the lowest growth recorded during the 301 days after transplantation (DAT) (Figure 2).

At six months after transplantation, the minimum and maximum height were 1.45 and 2.25 m, respectively. These values are lower than the 5.17 and 10.27 m reported by ^{Popoola et al. (2016}). This smaller size can be attributed to the lack of precipitation, since water has a direct effect on plant growth because in drought conditions, cell division and expansion decreases (^{Taiz and Zeiger, 2009}).

Figure 2 Plant height of the 20 accessions of Moringa oleifera collected in the South-Southeast of Mexico. Period: 0 to 301 DAT. 

In relation to the number of branches, there were few and these emerged after 2 m in height. ^{Popoola et al. (2016}) mentions that the branching is moderate. ^{Dao and Kabore (2015}) recorded in moringa, at two months, from 8 to 15 branches per tree. In this work, there was a monopodial growth in the accessions evaluated.

Stem diameter

The largest diameter at a height of 10 cm from the ground was observed in accession Y2 with an average value of 43.22 mm, and the lowest value corresponded to accession Y4 with 25.61 mm at 301 days after transplantation. The most frequent stem colors were gray and whitish gray. ^{Panshin and Zeeuw (1970}) mention that the thickening of the stem is due to related processes in xylem and phloem. Therefore, they may differ between accessions. The thickening of the stem and its morphology among moringa ecotypes is diverse due to the great plasticity it presents (^{Förster et al., 2015}).

Leaves

After transplantation, defoliation was observed in all accessions. At 301 DAT, accession Y3 had 54 leaves, the highest value being. Accession O1 presented 9 leaves, being the accession with the lowest average number of leaves. The shortest leaf length was 28.8 cm (Y3) and the longest was 47 cm (C2). The leaf width fluctuated between 16.8 (O1) and 38.33 cm (Y1). The presence of pubescence was recorded in accessions G4, C3 and G2. These values are higher than the range of 21.4 to 54.2 cm long and 10.1 to 41.6 cm wide reported by Zhigila et al. (2015). ^{Dao and Kabore (2015}) reported average leaf lengths from 16 to 44 cm and values from 10.5 to 34 cm wide, with a number of pinnas of 5 to 12 per leaf.

Leaf length is influenced by relative humidity and average annual precipitation. Phenotypic characteristics are affected by edaphic factors and result in epigenetic changes (Shahzad et al., 2013). Moringa leaves are consumed for their high nutritional value (^{Förster et al., 2015}). Therefore, the size, color and number of leaves represent an important characteristic to produce biomass, nutritional content and genetic improvement.

The presence of purple pigmentation in the petiole and rachis of accessions C1, O2 and Y2 was also identified. The dark color on the leaf represents a greater amount of chlorophyll and the existence of this photosynthate promotes greater growth (^{Opare-Obuobi, 2012}).

The violet pigmentation in the leaf petiole is determined by the anthocyanin content. The presence of anthocyanins is determined by environmental conditions and can be purple or pink in color. Usually, crops in areas under drought have higher anthocyanin production as a mechanism to avoid stress (Shahzad et al., 2013). The presence of pubescence on the leaves, tender shoots and filaments of the anther was noted. There was defoliation caused by high temperatures and little precipitation. ^{Vasconcelos et al. (2019}) mentioned that water stress caused by lack of precipitation influences leaf morphology and plant physiology. In this work, the effect of stress on the defoliation and yellowing of the leaves was observed.

Flowers

The beginning of flowering was recorded by counting the days since the sowing of the accessions. Accession O4 was the first to start the flowering process at 129 days after sowing (das). The last accession to flower was G3 at 240 das. Accessions Y2 and Y5 had pinkish-white flowers, C2 and V3 creamy white and O1 white flowers. Purple spots were identified in accessions G2 and V1 (Table 4). Flowering in cultivated accessions began at 129 days and continued during the months of October-May. However, the fall of flowers prevented many from reaching anthesis, limiting the pollination process.

Flowering in moringa can occur once or twice a year, depending on the environmental conditions (temperature and precipitation). ^{Price (2000}) mentioned that flowering can occur four times during the year. The pigmentation of the flowers varied from white, creamy white and pinkish white. Some accessions had purple pigmentation on the petals. ^{Popoola et al. (2016}) recorded white flowers with purple pigmentation, white or creamy white without pigmentation and 50% of flowering occurred in the 161 and 167 days. Moringa is a species that presents great variability in the color of its flowers and in some varieties of India, flowers with pink and dark pink base have been recorded (^{PPV and FR, 2001}).

Several studies have stated that pink and dark pink flowers receive more visits from bumblebees than yellow ones. This factor represents, apparently, a strategy of adaptation and reproduction (^{Bradshaw and Schemske, 2003}; ^{Reverté et al., 2016}). Therefore, cross-pollination between moringa accessions is promoted and facilitates the creation of new varieties. Factors such as temperature and soil moisture influence the increase of flowers, pollen viability and decrease the number of fruits (Muhl et al., 2013).

Pearson’s correlation coefficient (r) of morphological characters

The Pearson correlation matrix among the quantitative descriptors of moringa is shown in Table 5. A positive correlation was found between the length and width of the leaf (r= 0.857), and leaf length and petiole length (r= 0.851). A correlation was also found between diameter and height (r= 0.675). A negative correlation was identified between the days to first flowering and the number of flowers (r= - 0.599) and with the number of fruits (r= - 0.361).

Principal component analysis

The principal component analysis demonstrated variability in morphological characters in the 20 moringa accessions (Table 6). The percentage of variation was 52.87, 37.54, 6.16, 1.98 and 0.63 for components 1, 2, 3, 4 and 5, respectively.

The two-dimensional Figure 3 of the analyses of principal components 1 and 2 shows 3 groups. Group A comprises accessions O2, Y2, C2, G2, C3, V2, G4, C1, V3, Y1, O4 and V1. Group B comprises accessions G3, O3, G1, Y3 and C4. In group C, populations V4, O1 and Y4 were identified (Figure 3). The grouping was the result of the PCA based on the analysis of morphological descriptors.

Figure 3 Dispersion of components 1 and 2 for the 20 accessions of Moringa oleifera Lam., from the South-Southeast of Mexico. 

Cluster

Two subgroups were formed in group A: I (O2, Y2, C2, G2) and II (O4, G4, C1, Y1, V2, V3, C3 and V1). In group B, the subgroups: I (O3, G1, C4, G3) and II (Y3) were formed. In group C, accessions V4, O1 and Y4 were identified (Figure 4).

Figure 4 Cluster analysis based on morphological data from 20 accessions of Moringa oleifera collected in the South-Southeast of Mexico.