Multivariate Analysis of the Diversity among Some Nigerian Accessions of Amaranthus hybridus

MATERIALS AND METHODS

Sixteen accessions of A. hybridus were selected from the germplasm of the Nigerian Institute for Horticultural Research (NOHORT) based on germination and yield from previous field trials. The list of accessions and their seed colour (Table 1). The seeds were sown on seed beds in the Biological gardens of the University of Lagos in April, 1999. Twenty-eight days after planting, seedlings were transplanted into vegetable beds and planted at a spacing of 60x25 cm as recommended by Norman (1992). Animal (cow) dung was carefully mixed into each vegetable bed before transplanting.

Collection of data was based on 5 randomly selected plants per accessions and 5 measurements for each character trait per plant. A total of 22 characters were recorded in this study. They include leaf length and width, petiole length, internode length, stem diameter, plant height at maturity, plant canopy diameter, number of days to 1st foliage leaf, number of days to 4-leaf stage, number of days to 1st floral budding, number of days to 50% flowering, 500-seed weight, seed length and width. Based on these characters, three ratios leaf length/leaf width, plant height/plant canopy diameter and seed length/seed width ratios were determined. Visually assessed characters include colour of leaves, stem, inflorescence and seeds.

Means, standard deviation and coefficient of variation were used to measure central tendency and dispersion for the various quantitative characters recorded. Estimates of analysis of genetic diversity was carried out using SPSS for Window 11.0.

Table 1:	List of accessions used in the study and their seed colour

It involved calculating the degree of association among the different quantitative characters-Pearson’s correlation. The accessions were further classified into similar groups by the non-hierarchical k-means technique of cluster analysis with a dendogram (Hair et al., 1992).

RESULTS

Most of the quantitative characters measured exhibited broad variability. Plant height for example ranged from 74.30-181.00 cm, while days to 50% flowering had an interval of 23 days from the earliest to the latest maturing accessions. Most of the accessions were of green stem colour (81.25%) although some of these accessions (38.5%) had mixture of green and red colour. Only 1 of the 16 accessions had the stem colour being totally red (Table 2).

Of all the correlation coefficients 17 were positive and highly significantly correlated (p = 0.01) (Table 3). Most of the characters that were positively correlated corresponded to characters of morphology. The highest correlation was between seed length and seed breadth (R = 0.92).

Table 2:	Statistical parameters for quantitative characters and status of qualitative descriptors of the accessions evaluated

Table 3:	Correlations amongst selected quantitative traits in A. hybridus
LL = Leaf Length, LW = Leaf Width, PL = Petiole Length, IL = Internode Length, PH = Plant Height, PC = Plant Canopy, SD = Stem Diameter, FFL = Days to 1st Foliage leaf, 4LS = Days to four leaf stage, FFB = days to 1st floral buddings, 5OF = Days to 50% flowering, SL = Seed Length, SB = Seed Width, 500SW = 500 Seed Weight; **: Significant at p<0.01

Table 4:	Cluster composition of 16 Amaranthus accessions

Table 5:	Cluster means for some quantitative characters in A. hybridus accessions

Fig. 1:	Cluster analysis dendrogram of the 16 accessions of A. hybridus

Five hundred seed weight was also highly correlated with these two traits (R = 0.73 and 0.69), respectively. Generally, days to 50% flowering was negatively correlated with most traits although they were not significant. However, the value of R = -0.55 for plant height with days to 50% flowing was significant at p = 0.05.

Using the k-means non-hierarchical clustering the Amaranth accessions were grouped into 4 clusters (Table 4). There was only 1 accession in cluster 1 and plants in this accession had the highest values for morphological characters such as plant height (181.00 cm), leaf length (29.40 cm), petiole length (16.90 cm) and internode length (5.50 cm) (Table 5). They also flowered earliest in 86 days and it was a black seeded accession. The number of accessions in cluster 2 was six, five in cluster 3 and four in cluster 4. Accessions in cluster 4 had the lowest values for leaf length (17.33 cm) and plant height (80.78 cm). they also had the lowest value for seed characters, while the accessions in cluster 3 had the highest mean values for seed length and width which was 1.27 and 1.21 cm, respectively.

The results of the dendogram were consistent with the results in Table 4 as all accessions were recognized in the dendogram (Fig. 1). It was observed that the accessions coded 4 which was part of the accession in cluster No. 4 (Table 4) was separated from the other accessions in the dendogram.

DISCUSSION

The broad variability observed in leaf and seed morphology of the accessions studied is promising from the viewpoint of the genetic improvement of A. hybridus for leaf and grain vegetables. Wu et al. (2000) in a study of 229 genotypes from 20 Amaranthus species observed wide variability which was useful in cultivar improvement for agronomic traits, such as plant height, seed, stem and leaf colour among genotypes within same species and among different Amaranthus species. Similar results were also observed by Xiao et al. (2000) and Varalakshmi (2004), in the evaluation of different accessions of vegetable amaranths. Qualitative characters have also been opined to be important for plant description and are mainly influenced by the consumer’s preference, socio-economic scenario and natural selection (Zafar et al., 2006). They are also useful in separating varieties especially when the range of quantitative characters is limited (Ghafoor and Ahmad, 2003). Such characters in this study are represented by leaf and stem colour which make them acceptable to consumers. Most of the accessions in this study were of green leaf and stem colour which is a desirable qualitative character for selection especially when plant is used as a leaf vegetable.

Correlation measures the intensity of association between variables. In this study, the strong correlation between the leaf morphology variables and also between the seed morphology variables can be exploited directly/indirectly for the genetic improvement of the amaranths along these two lines-leaf and grain. The negative correlation between plant height and days to flowering could also be useful as the farmer can harvest for vegetables for a longer period before harvesting for grain especially if breeding lines could be developed that would combine high leaf and grain yields.

An important aspect of this study is the consistency between the dendogram and the cluster grouping of the accessions. In both cases, the accessions were in 4 distinct groups. This is obvious with the accession coded 3 which was the only accession in cluster 1 and was the most isolated accession in the dendogram. This accession which is quite different from all other accessions exhibited the highest value for leaf and plant height characteristics. Similar closeness was observed for accession coded 7, 8, 12, 15 and 16 which were in cluster 2 and in the group C of the dendogram. Although, in the dendogram, the accession coded 15 and 16 were the closest in the group, they all seem to be a subgroup/relation of the accession coded 7. Members in this group flowered earlier than accession in other groups.

From the results, accessions within each cluster seem to more closely related to each other than accessions in other clusters based on one or two characteristics. Accessions in cluster group 3 which were in group D in the dendogram had relatively high seed size values as measured by seed length and width while those in cluster 4/ Group B could be referred to as plants with small architecture as measured by morphological characters including seed size and weight.

Characterization of germplasm materials helps to ensure an efficient and effective use of such materials. Multivariate analysis provides an effective way of evaluating germplasm materials in order to identify materials that could be further evaluated and utilized at the genetic level. In this study, means, standard deviations, coefficient of variability, correlations, cluster analysis and dendogram have been used to assess the accessions and group them in a more reliable way using quantitative and qualitative characters. The quantitative characters have been useful as they tend to bring accessions that are genetically similar together in the multivariate analysis thus reducing the number of accessions in a breeding programme to a manageable number. Though, this is the fist attempt at classifying these materials along breeding lines, a few accessions can be selected for use directly in hybridization and also intervarietal hybrids can be developed. It is hoped that selection of accession with better mean performance from each group/cluster will give rise to better hybrids with the desired characters in future generations.

ACKNOWLEDGMENT

I wish to acknowledge the work done by Late Mr. F. Dibia on the Amaranthus accessions. He passed on before this write up was concluded.