MATERIALS AND METHODS

The present experiment was carried out in the coffee germplasm maintained at Regional Coffee Research Station, Chundale, Wayanad, Kerala, India during 2002-2006. The experimental field is located at an altitude of 840 m above mean sea level with an annual rainfall of 2000-3000 mm spread over a period of nine months ranging from March to November. Average humidity of the area is 88.9% with average minimum and maximum temperatures between 17.6 and 27.3°C. Soil is generally lateritic to laterites. The soil structure varies from sandy to clayey loams with the soil pH varying from 5.2 to 6.3. Seventy four robusta coffee accessions/genotypes, which include 61 robusta accessions identified from India and 13 exotic robusta accessions introduced by Central Coffee Research Institute, India from different coffee growing countries, planted during 1979-1983 period have been utilized for the present study (Table 1). All the plants studied were stabilized and mature during the period of data collection. The experiment was laid out in randomized block design, with three replications and fifteen plants per plot. The seedlings were planted at a spacing of 3x3 m and maintained uniformly as per the package of practices recommended by Coffee Board, India under rain fed conditions. Observations on 28 characters including 10 growth characters and 18 yield characters were made in 2004-05 when the plants became mature and stable in yield. The data were analysed for character association using the statistical software STATISTICA.

Table 1:	Accessions of robusta coffee studied for character association

RESULTS AND DISCUSSION

Five factors were obtained by factor analysis and the 28 characters under study could be grouped into five different groups as shown in Table 2 and 3. When analyzed based on relative factor loading, the characters showed that the first group consisted of 8 characters namely bush spread, length of primary branches, girth of primary branches, number of secondaries per primary, stem girth, leaf area, leaf length and yield per plant; the second group consisted of out turn (ripe to clean), bean breadth, bean volume, bean thickness, weight of 100 dry fruits, weight of 100 beans, bean length and weight of 100 fresh fruits; the third group consisted of internodal length, leaf breadth, out turn (dry to clean) and percentage of A grade beans; the fourth group consisted of out turn (fresh to dry), fruit length, fruit volume and fruit thickness and the fifth group consisted of number of primary branches, fruits per node and fruit breadth (Table 2, 3). Bush spread, length of primary branches and girth of primary branches were found to be the lead characters in the first group, out turn (ripe to clean) the lead character in the second group, internodal length the lead character in the third group, out turn (fresh to dry) the lead character in the fourth group and number of primary branches the lead character in the fifth group. Thus the study reveals the association of bush spread, length of primary branches and girth of primary branches with number of secondaries per primary, stem girth, leaf area, leaf length and yield per plant, the association of out turn (ripe to clean) with bean breadth, bean volume, bean thickness, weight of 100 dry fruits, weight of 100 beans, bean length and weight of 100 fresh fruits; the association of internodal length with leaf breadth, out turn (dry to clean) and percentage of A grade beans; the association of out turn (fresh to dry) with fruit length, fruit volume and fruit thickness and the association of number of primary branches with fruits per node and fruit breadth. This analysis show that bush spread, length of primary branches, girth of primary branches, number of primary branches, internodal length, out turn (ripe to clean) and out turn (fresh to dry) are the lead characters to be considered while planning breeding programmes in robusta coffee so that the bulk of variables for analysis could be reduced.

Table 2:	Factor analysis in the case of robusta coffee-factor loadings

Table 3:	Factor analysis in the case of robusta coffee-factors identified

Factor analysis can be used as an efficient tool to find out character association and to group the variables so as to effect data reduction by identifying the lead variables of each group. This method has been utilized already in crops like rubber (Abraham et al., 2002), cardamom (Radhakrishnan et al., 2004; Hrideek, 2007), tea (Ramasubramanian, 2005), rice (Mini, 2006) chillies (Hrideek et al., 2006) and coconut (Abdul Kadher et al., 2007).

The present study has shown that the polygenic agronomic characters of robusta coffee can be grouped in to five different groups based on factor analysis. This shows the association of characters in the species and provides important information on character groups that share common genetic factors. Moreover, the study has shown that bush spread, length of primary branches, girth of primary branches, number of primary branches, internodal length and out turn are the characters that are to be given premium importance while carrying out crop improvement programmes in robusta coffee.