Characterization of Some ex situ Conserved Finger Millet (Eleusine coracana (L.)) Germplasm Accessions in Sri Lanka

Finger millet (Eleusine coracana (L.) Gaertn.) is a highly nutritious and important food crop widely cultivated in the arid and semiarid regions in the world. Therefore it is worthy to be subjected to crop improvement programs. Germplasm collection and characterization are preliminary and important steps in crop improvement programmes. This study was conducted to characterize randomly selected 24 finger millet germplasm accessions conserved at the plant Genetic Resource Centre, Gannoruwa, Sri Lanka using 14 quantitative characters. The maximum positive and significant coefficient of variation was observed between weight of grain per ear and weight of sun dried ear. Phenotypic correlation between weight of grain per ear was highly significant and positively associated with days to flowering, flag leaf width, flag leaf length, plant height, culm thickness, finger length, finger width, days to maturity and weight of sun dried ears, flag leaf width, flag leaf length, plant height, culm thickness, finger length, finger width, days to maturity and weight of sun dried ears. The principal component analysis revealed that the first five component with Eigen values greater than 0.87 contributed about 85.5% of total variability. The twenty for finger millet accessions grouped in to four main clusters in the cluster analysis. Results of cluster analysis could be used in the crop breeding and conservation programmes.


INTRODUCTION
Finger millet (Eleusine coracana (L.) Gaertn.) is an allotetraploid (2n=4x=36) of the family Poaceae and subfamily, Chloridoideae and is commonly known as "Kurakkan" or "Kurahan" in Sri Lanka. It is an important food crop widely cultivated in the arid and semiarid regions of the world. It is native to Ethiopia and was introduced to India about 3000 years ago. Finger millet has been cultivated in Sri Lanka since ancient times and considered as the second staple food after rice in most of the rural areas. Finger millet is a highly nutritious crop with the highest amount of Calcium and Potassium content out of all cereal and millets. It contains a high amount of dietary fiber and have the ability of lowering blood glucose and cholesterol levels. Recently increasing attention is paid to improve finger millet due to above qualities and its inherent capacity to tolerate several abiotic stresses including water deficit and its adaptability to marginal soils with low fertility (Gana et al., 2013).
Germplasm is the basic raw material for any crop improvement programme. Characterization of germplasm accessions to identify true genetic diversity is also essential to use in crop improvement. About 200 finger millet germplasm accessions are being conserved at the seed gene bank of the Plant Genetic Resource Centre (PGRC), Gannoruwa, Sri Lanka. Some of those germplam accessions have been characterized by Kannagara et al., (2011) and Wakista et al., (2015aWakista et al., ( , 2015bWakista et al., ( , 2015cWakista et al., ( and 2015d using microsatellite markers and by Senanayaka et al., (2008) using AFLP markers. But largely those accessions are uncharacterized.
Multivariate statistical techniques are widely used in analysis of genetic diversity irrespective of whether it is morphological, biochemical or molecular marker-based. Among the multivariate techniques Principal Component analysis (PCA) and cluster analysis are very useful in germplasm characterization. Multivariate analysis has been used frequently in genetic diversity analysis many crops such as finger millet (Dagnachew et al., 2012, Ulaganathan & Nirmalakumari, 2015 and rice (Gana et al., 2013).
This study was undertaken to characterize some randomly selected conserved finger millet germplasm accessions using morphological markers to reveal their genetic relatedness.

Plant Material
Randomly selected 24 finger millet germplasm accessions obtained from the ex situ conserved accessions at the seed gene bank of Plant Genetic resource Centre (PGRC), Gannoruwa, Sri Lanka were used in this study (Table 1). These 24 accessions comprised of 19 local accessions collected from different geographical locations, two (02) Indian accessions, two (02) Zimbabwean accessions and two recommended Sri Lankan varieties. Seeds of the germplasm accessions were obtained from PGRC Gannoruwa.

Cluster analysis
Hierarchical clustering of complete linkage method with squared Euclidian distance was performed using MINITAB17 software. Data of all quantitative characters were standardized to a mean of zero and a variance of one before clustering to avoid bias that arise due to differences in measurement scales.

RESULTS & DISCUSSION
Twenty two finger millet germplasm accessions and two recommended finger millet varieties were characterized using 14 quantitative characters. Out of 14 quantitative characters one character (number of productive tillers) was monomorphic across all 24 accessions and other 13 characters were polymorphic. Wide range of genetic variability was observed for different characters and their mean, minimum and maximum values are given in Table 2.
Earliness is an important agronomic trait considered while breeding for high yielding varieties.

SD -Standard Deviation
Knowledge on the association between yield and other biometrical traits provide ample opportunities for improvement of crop. The correlation between characters may exist due to various reasons. The association of all quantitative characters was estimated by correlation analysis (Table 3). Out of all studied quantitative characters, days to flowering (days to flowering ( Grain yield per plant was equal to the weight of grain per ear with respect to all accessions used in this study as there were no productive tillers. Significant negative correlation (at ≤ 0.005 level) observed for days to flowering with clum branching (-0.436) indicated increasing one character would lead to decrease in another character. The character 1000-grain weight did not show any significant correlation with any of other characters.      Similarly, the recorded characters were divided into two groups (X and Y) depending on the first two principal components as shown in PCA loading plot (Figure 2).
The group X consists of Days to flowering, days to maturity, Finger length, Finger width, weight of sun dried ear, weight of grain per ear and 1000 grain weight whereas Plant height, Flag leaf length, flag leaf width, culm thickness, culm branching and finger number had been grouped in the group Y.

CONCLUSIONS
Characterization of available germplasm accessions is very important in crop improvement programmes and morphological markers play an important role in germplasm characterization.
Fourteen quantitative characters were used in the study and Principal Component Analysis revealed that days to flowering, flag leaf width, flag leaf length, plant height, culm thickness, finger length, finger width, days to maturity and weight of sun dried ears had significant and positive correlation with