Tanzania Journal of Agricultural Sciences 2003, Vo1(1) : 45-54The present investigation evaluates the performance of newly identified clones in the major coffee growing zones of Kagera region, Tanzania. Five clones namely MS1/95. MS2/95. MS3/95, MS5/95. MS6/95 selected from individual trees in farmers fields and FS a control variety, were grown in farmers fields in March 1998 a four locations representing 3 coffee growing zones of Kagera region. Each chosen farmer's field was a replicate. Plant girth and fruit set percentage differed significantly among the clones tested MS2/95 had the thickest stems while MS3/95 gave the highest percentage of fruit set. Percent fruit set ranged from 13.5% - 31.4% for MS6/95 and MS3/95 respectively. Locations differed on yield % bearing primary branches canopy radius plant girth and plant height. Kabirizi B gave the highest yield of 1853.8 kg/ha while Chanika gave the lowest yield of 358.8 kg/ha. The yield differences between location were associated with performance of clones on plant height, plant girth canopy radius, primary branches and % bearing primary branches. Plant height girth canopy radius, primary branches, % bearing primary branches and yield of clean coffee were positively corrrelated among themselves. Heritability estimates were high (50%) for plant girth, berries per node and fruit set percentage. Heritability ranged from negligible for yield to 125% for plant girth. Expected genetic gain varied with heritability ranging from negligible to 36.5% for fruit set percentage. Yield of clean coffee had appreciable amounts of clone x location interaction and environmental variance in relation to the total phenotypic variance. It should be possible to realize substantial gain from selection and genetic improvement for plant girth, berries per node and fruit set percentage.