MATERIALS AND METHODS

The experiment was carried out in the experimental farm of Indian Cardamom Research Institute (ICRI), Myladumpara, Idukki, Kerala, India during 2002-2006. The farm is located at an altitude of 1,068 m above MSL at 9° 53N latitude and 77° 09 E longitude and it enjoys humid tropical monsoon climate. The soil is forest loam with pH of 5-6. The experiment was laid out in randomized block design with three replications and 12 plants per plot at a spacing of 3x3 m. Eleven genotypes consisting of ten elite native landraces and one released variety were used for the study (Table 1). Package of practices recommendation of the Spices Board, India was followed for cultivation. Observations on seven growth characters, nine yield attributes and eleven quality parameters were recorded and analyzed.

Biometric study of growth, yield and quality characters of the eleven genotypes has been used to analyze character association in cardamom. Factor analysis by means of principal component analysis (Sneath and Sokal, 1973) was carried out for the purpose with the help of the software STATISTICA.

Table 1:	Cardamom genotypes used for the study

RESULTS AND DISCUSSION

The seven growth characters of cardamom studied presently could be grouped in to two based on factor loading with number of bearing tillers showing maximum factor loading in the first group and leaves per tiller showing maximum factor loading in the second group (Table 2-4). Out of the two factors identified among the growth characters the first factor with four characters contributed 33.79% of variance exhibited by the growth characters studied and the second factor consisting of three characters contributed 31.08% of variance. Characters with the highest factor loadings can be considered as lead characters based on which selection could be practiced. This analysis has shown that number of bearing tillers and leaves per tiller are the most important characters to be considered when selection is practiced for growth characters of cardamom. It also indicated that if selection is carried out based on the above two characters, the other characters will automatically get selected favorably.

Table 2:	Factor analysis of growth characters in the eleven cultivars of cardamom studied- percentage of variance contributed by each factor and cumulative percentage of variance

Table 3:	Factor analysis of growth characters in the eleven cultivars of cardamom studied- factor loadings

Table 4:	Factor analysis of growth characters in the eleven cultivars of cardamom studied- factor groups and the characters in each group

Table 5:	Factor analysis of yield characters in the eleven cultivars of cardamom studied- percentage of variance contributed by each factor and cumulative percentage of variance

Table 6:	Factor analysis of yield characters in the eleven cultivars of cardamom studied- factor loadings

Table 7:	Factor analysis of yield characters in the eleven cultivars of cardamom studied- factor groups and the characters in each group

Table 8:	Factor analysis of quality characters in the eleven cultivars of cardamom studied- percentage of variance contributed by each factor and cumulative percentage of variance

Table 9:	Factor analysis of quality characters in the eleven cultivars of cardamom studied- factor loadings

Table 10:	Factor analysis of quality characters in the eleven cultivars of cardamom studied- factor groups and the characters in each group

The yield characters under study could be grouped into two factors based on factor loading and the first group consists of five characters and the second group consists of four characters (Table 5, 6 and 7). Seeds per capsule could not be grouped in any of these groups. The first factor contributed 34.9% of variation contributed by yield characters and second factor contributed 23.68% of variability. Dry yield per plant was found to be the lead character in the case of the first group of variables and internodal length was found to be the lead character in the second group of variables.

Factor analysis of eleven quality characters resulted in the grouping of the characters in to three groups based on factor loading (Table 8-10). The first group consisted of five characters, second group consisted of three characters and the third group consisted of one character. However, the rest of the variables could not be grouped in to any of these factor groups. The three factors together contributed 74.80% of cumulative variance. The percentage of variance contributed by the first factor is 40.71, that contributed by second factor is 20.64 and that contributed by the third factor is 13.46. The lead character in the case of the first group is moisture content followed by percentage of bold (7 mm and above) capsules. In the second factor group, volatile oil content is the lead character and in the third factor group 100 capsule weight-dry is the only character involved.

Factor analysis is a very efficient tool used to find out character association and to group the variables in to different groups and also to effect data reduction by identifying the lead variables of each group and the technique has been utilized by earlier workers in crops like rubber (Abraham et al., 2002), coconut (Abdul Kadher et al., 2007) and chillies (Hrideek et al., 2006). A study of character association in cardamom by Sritharan et al. (1993) revealed that number of pseudostems per plant, leaf length, number of panicles per plant, panicle length and number of capsules in a panicle had higher association with yield than other parameters. Pooled factor analysis resulted in the grouping of 17 characters of cardamom in to six factors in an experiment carried out by Radhakrishnan et al. (2006).

The present study has revealed that the quantitative agronomic characters of small cardamom can be grouped in to three groups based on factor analysis. Moisture content in capsules, percentage of bold capsules, volatile oil content and weight of capsules have been identified as the lead characters in small cardamom based on which further improvement programmes can be designed.