Morphological Variation in Arabica Coffee (Coffea Arabica L.) Growing in North Sumatra Indonesia

Genetic variation is important in plant breeding. However, information on the genetic variability of Arabica coffee especially in coffee field of North Sumatra was not yet available. Magnitude of morphological variation, genotypic variation, phenotypic variation, heritability, genetic advance, genetic correlation, and phenotypic correlation of plant vigors and yield components of 28 genotypes were evaluated using nested design. This research showed morphological and genetic variations of the genotypes in the field. Based on the research locations as operational taxonomic unit, the genotypes were separated into three clusters. Most of the parameters had low to moderate genotypic variation, while phenotypic variation was moderate to high. Heritability and genetic advance were low, moderate, and high. Several plant vigors and yield components had a positive significant genetic and phenotypic correlation one another, and several had negative ones. Coffee berry borer infestation (CBBI) had a highly significant negative genetic correlation with leaf width (rG = -0.309**), leaf weight (rG = -0.671**), fruit diameter (rG = -0.320**), and bean length (rG = -0.175**). CBBI showed a significant positive genetic correlation with mesocarp pH (rG = 0.134*). To reduce CBBI, selection for higher leaf weight is better. Selection on lower pH of mesocarp could be considered to decrease CBBI.


INTRODUCTION
Arabica coffee (Coffea arabica L.) was cultivated for the first time around 120 years ago in North Sumatra. In Indonesia, North Sumatra is one of the most important production center of Arabica coffee. Indonesia produced Desember 2018 pressures might create genetic mutation lead to genetic variation.
Although Indonesian Goverment has released several commercial cultivars, empirical facts showed that many of coffee farmers are still using traditional seeds from unknown resources for their new cultivation field which might cause low coffee productivity (1.14 ton ha -1 of green bean), and might cause genetic variation among farmers' land. Low productivity could also be affected by soil fertility (Hanisch et al., 2011) and coffee berry borer (CBB) attact which is considered as one of the most destructive pest of Arabica coffee in North Sumatra.
Plant breeders require genetic variability of desirable characters to carry out the breeding programs (Mayo, 1987;Mishra and Slater, 2012;Constantin et al., 2017). Genetic variation of Arabica coffee can be found not only in cultivated cultivars (Setotaw et al., 2010;Tessema et al., 2011;Geleta et al., 2012;Fatimah et al., 2014;Randriani et al., 2014;Dani et al., 2016), but also in wild populations (Schmitt et al., 2009;Aerts et al., 2013;Atinafu et al., 2017). Previous studies done by Silvestrini et al. (2008) and Kathurima et al. (2012) exhibited that genetic variation in commercial coffee cultivar was narrow. However, another study conducted by Geleta et al. (2012) revealed broad genetic variation in the collection of Arabica coffee cultivars. Genetic diversity was shown to be correlated with morphological diversity (Yuan et al., 2015). However, information on genetic diversity of Arabica coffee derived from North Sumatra was not yet available. The aim of this research was to determine morphological and genetic variations of Arabica coffee at coffee plantations in North Sumatra.

MATERIALS AND METHODS
The research was carried out in District Tapanuli Utara, Toba Samosir, Humbang Hasundutan, Samosir, Simalungun, Pakpak Bharat and Dairi, North Sumatra Province. Data was collected in July 2014. The nested design with three factors was used for data analysis (Quinn and Keough, 2002). The first step was to select 7 districts, then 2 sub-districts were chosen in each district, and the final step was to select 2 coffee farms in each sub-district. These selected coffee farms were treated as genotypes (G). Each farm consisted of 200-300 plants of variety Sigarar Utang which is Arabica coffee. The plants were 6-7 years old, with the characteristics of having a shot of bronze-colored leaves, ripe fruits, and harvest frequency of once in two weeks. Ten plants were selected randomly in each farm. In total, twenty eight genotypes of Arabica coffee were used to determine morphological and genetic variation in this study. Mesocarp pH was measured using pH meter (Amtast KS-05 vergara). A fruit showing the frass on the entrance hole is a CBB infected fruit which is caused by females of CBB live inside the fruit after boring a hole at dictus or near the dictus (Vega et al., 2009). All fruits were checked. Coffee berry borer infestation (CBBI) was the ratio of the number of infected fruits to the total number of fruits (%).
Tree morphology comprises plant vigor (plant height, leaf length, leaf width, leaf weight), yield components (100 fruits weight, fruit length, fruit diameter, mesocarp thickness, mesocarp pH, 100 parchments weight, parchment length, parchment width, parchment thickness, 100 beans weight, bean length, bean width, bean thickness), and coffee berry borer infestation (CBBI) (Wahyudi et al., 2016). All data were analyzed with the hierarchical cluster analysis using nearest neighbour cluster method measured with squared Euclidean distance. In the analysis, research location was used as operational taxonomic unit (OTU) while coffee morphology were treated as variables.
XVLQJ WKH GHJUHH RI IUHHGRP RI WKH HUURU 4XLQQ and Keough, 2002). IBM SPSS version 19 and Microsoft Excel version 2007 was used for data analysis.

RESULTS AND DISCUSSION
The cluster dendrogram showed morphological variation among research locations ( Figure 1). Gichimu and Omondi (2010) found the correlation between morphological variation with genetic variation of coffee genotypes. Genotypes were significantly different in plant vigor, yield components, and CBBI (Table 2). This research found low and moderate genetic variation in several plant vigor and yield components while high one in CBBI (Table 3). The results of this research might be generally in line with Kitila et al. (2011) and Beksisa and Ayano (2016) who found low, moderate and high genetic variation in fruit length and fruit diameter, plant height, bean length, and bean width. Tessema et al. (2011) found the similar result in bean weight but Kitila et al. (2011) revealed high genetic variation in bean weight. Low and moderate genotypic variation in most of the parameters might indicate the nature of self-fertilized coffee plants. Broad genetic variability must be obtained through hybridization.

Desember 2018
Plant vigor and yield components showed moderate to high phenotypic variation (Table 3). The results of this research supported Kittila et al. (2011) and Beksisa and Ayano (2016) who found moderate to high phenotypic variation in plant height, fruit diameter, bean length and bean width. In contrary to this research, low phenotypic variation in fruit length and high phenotypic variation in bean weight were found by Kitila et al. (2011) and Tessema et al. (2011). This research showed low, moderate, and high heritability in several plant vigor component and yield components (Table 3). High heritability was manifested by leaf weight, fruit weight, mesocarp pH, parchment weight, and CBBI. These research results were in line with Kitila et al. (2011) who found high heritability in plant height, fruit length, fruit diameter, bean weight, bean length and bean width. However, Kitila et al. (2011) found high heritability in plant height, fruit length, fruit diameter, bean weight, bean length and bean width while Bekisa and Ayono (2016) revealed low heritability in plant height, and Tessema et al. (2011) found high heritability in bean weight. This research revealed low to high genetic advance in several plant vigor components and yield components (Table 3). In contrary to this result, Kitila et al. (2011) and Bekisa and Ayono (2016) found moderate genetic advance in several plant vigor and yield components. Kitila et al. (2011) and Tessema et al. (2011) found high genetic advance bean weight. Two-thirds (66.7%) of the parameters had low to moderate genetic advance. This might relate to narrow genotypic variation of the existing coffee cultivars as selffertilized coffee plants.
This research found that all plant vigor components (plan height, leaf length, leaf width, leaf weight) had a high significant genetic correlation (Table 4). Genetic correlation between several vigor parameters one another and with yield components was also found by Kitila et al. (2011). Selection for leaf weight would be the first priority to increase resistance of plant against CBB. The selection could be possible to be carried out successfully due to high heritability. The lower pH of mesocarp was the less CBBI was. Consequently, selection for lower pH of mesocarp could decrease CBBI. The selection could be conducted successfully due to moderate genetic variation. This selection may be combined with selection for yield in the first high-yield year (Oliveira et al., 2010). This research revealed that several plant vigor and yield components phenotypically correlated each others (Table  4). Kitila et al. (2011), Rodrigues et al. (2012), and Gessese et al. (2015) found a phenotypic correlation between several plant vigor parameters and yield components.

CONCLUSION
This research revealed morphological and genetic variation of the genotypes of Arabica coffee. The genotypes morphologically separated in three clusters based on the research locations. Leaf length, leaf width and leaf weight, hundred fruit weight, mesocarp thickness, mesocarp pH and hundred parchment weight showed moderate genetic variation. Plant height, fruit length, fruit diameter, parchment length, parchment width, parchment thickness, hundred bean weight, bean length, bean width and bean thickness had low genetic variation. Because leaf weight had significant negative genetic correlation with coffee berry borer infestation, selection for higher leaf weight Desember 2018 would be the best selection criterion to improve resistance of coffee against coffee berry borer. In future research, it could be needed to examine how pH of mesocarp could affect CBBI. Coffee hybridization is needed to obtain broad genetic diversity and big genetic advance.