Correlation between seed characteristics and biomass production of Moringa oleifera provenances grown in Ouagadougou , Burkina Faso

The correlation between seed size and biomass production of Moringa oleifera, was conducted in a field experiment at Ouagadougou (Burkina Faso) in 2014. The experiment included eleven M. oleifera provenances from West Africa according to four agro-ecological characteristics (Sahelian, Sub Equatorial, South Sudanian and North Sudanian). Differences among provenances seed traits, leaf morphology, and dry mass (DM) were established by analyses of variance. Significant level was fixed at P < 0.05. Pearson’s correlation analysis was done to evaluate the relationships between seed traits, leaf morphology, and dry mass. The results showed that significant differences in seed traits were observed among provenances. Provenances were significantly different in leaf morphology (length, width and number of pinnae/seedling). The provenance of National Forest Seed Center (CNSF) from the northern sudanese area of Burkina Faso produced the longest leaf. The mean shoot biomass accumulation per seedling differed significantly among provenances. Significant correlation coefficients were observed among seed traits but no important correlation were found between seed traits and other plant characteristics. No significantly correlation of seed size with either leaf morphology or dry mass were found. Plant grown from large seeds compared to those grown from small seeds was more vigorous and produces greater dry matter.


INTRODUCTION
There are many trees and shrubs of interest in agroforestry systems, and one interesting tree species that has received a great deal of attention recently is Moringa oleifera Lam (Mendieta-Araica et al., 2013).The species is planted to produce pods, leaves and seeds for consumers, forage and traditional medicinal purposes.The pods and leaves of M. oleifera contains high amount of Ca, Mg, K, Mn, P, Zn, Na, Cu and Fe (Aslam et al., 2005).
In Burkina Faso, M. oleifera leaves are commonly used *Corresponding author.E-mail: mamadjelia@yahoo.com.
Author(s) agree that this article remains permanently open access under the terms of the Creative Commons Attribution License 4.0 International License to make sauces and as staple food.Cultivations of M. oleifera within different environmental conditions throughout the year attract attention of poor smallholders during the last years (Popoola and Obembe, 2013).Biomass productions and seeds yield are of key economic importance.Some smallholders in Burkina Faso formed the associations of women nursery operators engaged in seedlings productions for income generation with the objective of selling the seedlings.They have been found to exchange most of the M. oleifera seeds often collected from a few mother isolated trees in their farmlands, and plantations in the communities or outside.This way of supplying seeds by private producers limited the access to high quality germplasm (Nyoka et al., 2015).M. oleifera plants were promoted within smallholders by many non governmental organisations (NGOs) and government, but no varieties with desirable traits for specifics growing conditions (home gardens, farmlands) are available, which makes its cultivation a risky business.
Good seedling establishment is essential for sustainable and profitable crop production, and is therefore recognized as the most critical step of a developing plant.Low seed vigor greatly controls the number of emerging seedlings, the timing and uniformity of seedling growth.This has direct influence on the yield and marketing quality of a crop (Pallo et al., 2009).
Seed weight is an indication of the reserves that seeds contain, and large and heavy seeds reveal that the seed has more reserved food (Wolde-Mieskel and Sinclair, 2000).Many studies have shown that initial seedling size is positively related to seed size, and larger seeds have better seedling survival rate as well as higher competitiveness both within species and among species (Singh et al., 2006).The seed supplies the embryo with sufficient nutrition and energy during germination from the food reserves that the seed acquires during the seed filling phase.Roxas et al. (1994) concluded that higher vigor that occurred in larger seed is due to the larger food reserves in these seeds.
Many studies also indicated a positive linear relationship between seed weight and emergence in the field.Baalbaki and Copeland (1997) reported that in wheat, seed size not only influence emergence and establishment but also affected yield components and ultimately grain yield.A similar observation was made by Arunachalam et al. (2003), while working with the tree species, and this was attributed to the larger food reserves in the larger seeds.
Several studies have been conducted looking for M. oleifera growth performance and biomass production (Edward et al., 2014;Förster et al., 2015), biomass production and chemical composition (Mendieta-Araica et al., 2013), seed and germination characteristics.There are however limited studies on the explicit relationship among seed and biomass characteristics.
In addition, the multivariate analyses have been widely used to study the influence of interacting traits on fitness in plant ecology studies (Farris and Lechowicz, 1990), assess plant responses to environmental stress (Hofmann et al., 2013), and examine morphological variation in agricultural crops (Ayana and Bekele, 1999).
The objective of this study was to investigate and compare different growth characters among provenances of M. oleifera in smallholders gardening systems.Seed length, width weight, leaf size and shape, and branch number may be useful predictors of potential production in specific environments (Pellis et al., 2004;Weih and Nordh, 2005).The relationships between seed characteristics and biomass production of M. oleifera in north sudanian part of Burkina Faso were therefore investigated.

Location of experimental area
The study was carried out at the women gardening center named "Amicale des Forestières du Burkina Faso (AMIFOB)" located at Ouagadougou, Burkina Faso (12°7"32""N, 01°40"24""W).This corresponds to an ecological zone of the north sudanian forest, with average annual rainfall of 800 mm.The soils are sandy clay to claysandy Ferruginous leached with very low nutrient content according to French soil classification (Pallo e al., 2009).The common natural vegetation found at Ouagadougou is described as semi-deciduous open woodland.Main genera include, Eucalyptus, Azadirachta, Mangifera, Vitellaria, Lannea, Piliostigma, Acacia., Ziziphus, Tamarindus, and Combretum.

Experimental design and sampling procedures
Trees were planted in a randomized complete block design (RCBD) with three replications.Each plot represented a provenance planted at 5 x 6 rows in a contiguous arrangement of 20 x 20 cm (Figure 2).Plot measured 10 x 1 m and contained 30 trees, and the distances between blocks were 2 m.Seed samples were pretreated with water for 24 h, and sown in June 1st 2014 in a prepared and cleaned soil using hand hoes.Weeding was done twice during the rainy season.Watering was done once a day in the morning.The seedlings were grown without fertilizers and chemical control.

Data collection and analysis
To investigate the variability in seed parameters (length, thickness,  and weight), each provenance was represented by 240 randomly selected seeds, assessed in four replications of 60 seeds each.Each seed weight was determined by weighing three random samples of 60 seeds each.By the end of 2 weeks; the experiment had been measured twice at effective ages of 5 and 12 days since seed sowing, providing data for the assessment of germination rates of the twelve provenances.Total samples of 20 seedlings per replication for each provenance were assessed.The seedlings were separated into shoot and root components that were ovendried for 48 h at 70°C, and weighed to determine total dry weight and calculation of root-to-shoot ratios.Height was measured using calibrated height measuring pole.Differences among provenances seed traits, leaf morphology, and dry mass (DM) were established by analyses of variance (ANOVA) using the Linear Model procedure (Standard Least Squares option) of the JMP® Pro 12.0.1 statistical software package (SAS Institute Inc., Cary, NC).Normality and homoscedasticity were graphically verified on residual plots of the linear models (Quinn and Keough, 2002).Where the ANOVA indicated significant treatment effects, treatment means were separated by Tukey"s Honestly Significant Difference (HSD) test (p = 0.05).Pearson"s productmoment correlation analysis was done to evaluate the relationships between the studied traits (Seed weight, distance, thickness; Leaf length, width, pinna number; base, shoot, total, root/shoot dry mass).The provenance relationships (""similarity"") were examined on the basis of the 10 traits used in the correlation analysis by a hierarchical cluster analysis using the JMP® Pro Cluster Analysis from the Multivariate methods submenu, which applies the ""Ward distance"" (SAS Institute Inc., 2012).
Mean number of pinnae per seedling at the same time ranged from 6 to 12 for Segou and CNSF, respectively.At the end of two months, the mean shoot biomass accumulation per seedling differed significantly (P < 0.05) among provenances (Table 2).The lowest shoot dry weight of 2.2 g was collected in provenance Bobo Dioulasso, while the highest of 21.4 g was obtained in provenance Gaoua.Provenance differences in root dry biomass per seedling were considerable, with means ranging from 1.2 for 7.4 g for Tamalé and Gaoua, respectively (Table 2).Mean shoot ranged from 1.2 to 14.0 g for Bobo Dioulasso and Gaoua, respectively (Table 2).Similarly, total dry mass produced at the same time ranged from 2.2 to 21.4 g for Bobo Dioulasso and Gaoua, respectively.The ratio root: shoot dry mass ranged from 0.4 to 0.9 g for Dano and Tamalé, respectively.

Interdependence among characters
Significant positive Pearson"s product-moment correlation coefficients (r) were observed among seed traits but no important correlation were found between seed traits and other plant characteristics (Table 3).Traits related to seed size (length, width, and weight) were mostly correlated with each other, Seed weight was positively correlated with seed length (r = 0.28), and seed thickness (r = 0.3).Seed length was positively related with seed thickness (r = 0.23).No significantly correlation of seed size with either leaf morphology or dry mass.An exception to the trend was seed thickness with leaf length, shoot dry biomass and total dry mass, which displayed a weak magnitude of the correlation coefficients (r < 0.20).Significant relationships were also detected between traits related to leaves morphology,    shoot and root dry biomasses.Leaf length was highly correlated with leaf width and number of pinnae (r > 0.80) (Table 3).Leaf morphology was positively correlated with the aforementioned while root dry mass and total dry mass (r > 0.40) were negatively correlated with root: shoot dry biomass (Table 3).

Cluster analysis
Cluster analysis based on seed sizes, leaf morphology and biomass production revealed two distinct clusters (Figures 3 and 4).Provenances Ouahigouya, Segou, Bobo-Dioulasso and Tamalé located in three different ecological zones formed cluster 1 while the seven other provenances (Dano, Ouagadougou, Fada N"Gourma, Gaoua, Koudougou, CNSF and Abidjan), dispersed also in three different ecological zones formed the second cluster group.Provenances from cluster 1 performed high seed traits, small leaves sizes and lower biomass accumulation than cluster 2. Tamale and Bobo Dioulasso in particular, invested more biomass in root than in shoot (Figure 4).Cluster 2 had large leaves size and excellent dry biomass production.

Seed
Seed size is the main determinant of maternal investment in individual seed farmer, and it as been demonstrated that tree-to-tree variation in seed size within local populations is under relatively strong genetic.The relationship between seed size and the growth of resulting seedlings can last for about one year or up to 10 years (Farmer, 1997).Investigation of DNA markers have demonstrated that there is a large genetic diversity present in wild collections of Moringa oleifera (Shahzad et al., 2013).The three seed traits differ significantly between provenance, specially the seed weight.Difference between the heaviest (Gaoua) and the lightest (Koudougo) was up to 43%.No environemental gradient, longitude, latitude, altitude, and annual rainfall explain difference in seed traits between provenance.Grouping provenance in ecological zones did not explain either diffence in seed traits.In the opposite, the regional provenances of Faidherbia albida showed a consistent variation in seed length, seed width and seed weight: the southern African provenances had the largest seeds and west African provenances the smallest (Dangasuk et al., 1997).
According to these authors, seed weight and size are adaptive traits, their smaller mean values observed in the west African provenances might suggest strong genetic selection for small seed size which is probably an adaptation to the greater desiccation stress in the Sahelian region.In their study of provenance variation in seed size of Senegal mahogany, Ky-Dembele et al. (2014) indicate that genetic differences exist among families within provenances and provenances in seed traits of K. senegalensis in Burkina Faso.In contrast, the high inter-and intra-provenance variation in Moringa seed weight in this study might suggest that this parameter is quite insensitive to environmental factors.
As Moringa seed weight shows significant differences between years (Ayerza, 2011), difference in seed weight could be attributed to the year's harvest.Seed weight has little effect on acacia seedling growth in different provenances of Senegalese (Wolde-Mieskel and Sinclair, 2000), Singh et al. (2006) have found a positive correlation between morphological characters of Celtis Australis seeds, family Ulmaceae, including seed weight and elevational gradient.Provenances with heavier seed may provide seedlings that will be able to withstand the adverse climatic conditions.

Growth performance
The study results agree with those of Gamedze et al. (2012) that shows significant differences in growth performances between Moringa provenances in Swaziland.The three provenances had significant differences in the number of leaflets produced.

Conclusion
It may be concluded from this study that significant differences were found among seed traits, leaf morphology, dry diomass and provenances.The research revealed that seed size were not correlated with either leaf morphology or dry mass.However, additional studies are needed to determine factors affecting moringa yields, including potential interactions of genotype×environment on seed per tree production, before making any recommendation about the economic potential of moringa as a new crop for this region.

Figure 3 .
Figure 3. Hierarchical cluster dendrogram of the 11 M. oleifera provenances based on seed traits, leaf characteristics, and dry mass.

Figure 4 .
Figure 4. Mean ± standard error of seed traits, leaf characteristics, and dry mass (DM).Provenances are grouped according to the hierarchical cluster analysis.

Table 1 .
Location and other attributes of the M. oleifera provenances. Prov.

Table 2 .
Mean, range and standard-error of seed traits (length, thickness, weight), leaf morphology (length, width, pinna number) and dry mass (Root, shoot, total, Root/Shoot) of the 11 provenances of M. oleifera.

Table 3 .
Pearson correlations between seed traits, leaf morphology and dry mass.