This study was undertaken aiming to determine the existence of linear correlations, based on simple regression studies for a better improvement of young rubber tree (Hevea spp.) breeding and selection. The characters studied were: yield of dry rubber per tapping by Hamaker-Morris-Mann test tapping (P), mean gurth (CC), bark thickness (EC), number of latex vessel rings (NA), diameter of latex vesseis (DV), density of latex vesseis per 5mm within rings (D) and distance between latex vesseis rings (DMEAVC) in a three and haif years old nursery. The results showed that among other characters, the correlation values between P and CC, EC, NA, DV, D, DMEAVC were r = 0.61, 0.34, 0.28, 0.29, 0.43 and -0.13 respectively. The correlation coefficients values between CC and EC, NA, DV, D, DMEAVC were: r = 0.65, 0.22, 0.37, 0.33 and 0.096 respectively. Linear simple regression of P with CC, EC, NA, DV, D and DMEAVC suggested that CC was the only important and significant character accounting for 36% of the variation on P. Concerning with vigor, the regression of CC on the same characters suggest that EC was the only important and significant character accounting for 42% of the variation on CC. The high correlation of mean girth with yield and bark thickness showed that it is possible to obtain high yielding young ortets through early selection of these characters.
rubber tree; Hevea spp.; bark anatomic characters; linear correlation and regression