Ranking Traditional Rice Cultivars Based on Yield and Some Morphological Traits using Path Analysis, Multi-Criteria Decision Making Model and Compromise Programming

Aims: To rank traditional rice cultivars based on yield and some morphological traits.
Study Design: Experiment was conducted with four replications according to the randomized complete block design. Germinated seeds were planted in rows with 15 cm X 20 cm spacing with twenty plants per each line and three lines per each replication. Data were collected in middle raw of each three lines in four replicates.
Place and Duration of Study: A field experiment was carried out during 2011/2012 Maha and 2012 Yala seasons at Faculty of Agriculture, University of Ruhuna, Sri Lanka.
Methodology: Plant height (cm), number of tillers/plant and number of fertile tillers/plant were counted before harvesting and panicle length (cm), panicle weight (g), number of spikelets/panicle and number of fertile spikelets/panicle were measured after harvesting in 80 plants of each rice cultivar in four replicates. Hundred grain weight (g) and yield/plant (g) were measured after drying grains for 14% moisture content. Total effect of each yield attributing character to the final grain yield which is defined by total effect of path analysis was calculated using SPSS software. Traditional rice cultivars were ranked according to yield and agronomic characters by multi criteria analysis and compromise programming technique.
Results: According to the path coefficients, number of fertile spikelets/panicle (β=0.982) and number of fertile tillers/plant (β=0.787) recorded the highest positive direct effect on grain yield. Plant height (β = 0.036), panicle length (β=0.048), panicle weight (β=0.305) 100-grain weight (β = 0.260) and filled grain percentage (β=0.130) recorded positive direct effects on grain yield. Negative direct effects were recorded for the number of tillers (β -0.267), number of spikelets per panicle (β =-0.363). Plant height (β =0.179), number of tillers/plant (β =0.838), number of fertile tillers/plant (β =0.077), panicle length (β =0.034) and number of spikelets/panicle (β =0.733) had positive indirect effects on grain yield while number of fertile spikelets/panicle (β=-0.122) recorded negative indirect effects on grain yield. Hence total effect of individual trait on yield was considered for ranking traditional rice cultivars in muti-criteria decision making model where number of fertile spikelets/panicle (0.870), number of fertile tillers/plant (0.864), plant height (0.571), panicle length (0.082), panicle weight (0.305), 100 grain weight (0.260) and filled grain percentage (0.141) effect differently on grain yield. All the rice cultivars were ranked from 1 to 100 according to yield and agronomic characters using total effect of path analysis.
Conclusion: Among tested hundred traditional rice cultivars Hondarawala was the best genotype followed by cultivar EAT Samba and cultivar Bathkiri el. All the other cultivars were placed in between 2-99 positions according to the results of multi criteria decision making model.