【Objective】 Clonal testing plays a pivotal role in forest tree breeding programs by enabling the evaluation of genetic variation among clones and facilitating accurate estimation of genetic gain from selection, while few studies have been conducted on the second or higher rotations of clonal test in eucalypts. This study aimed to quantify the genetic variation and clonal repeatability of growth traits at 3.5 years in the second rotation of eucalypt clones, and evaluate the effectiveness of across-rotation selection strategies for identifying superior clones. 【Method】 A set of cutting rooted full sib clones derived from a controlled cross between Eucalyptus urophylla and E. tereticornis were used for estimating clonal repeatability (H²) of 3.5-year-old growth traits of the second rotation as well as trait-trait phenotypic correlation (r_p) and genetic correlation (r_g) for these growth traits compared with these important traits of the first rotation. Clonal breeding values (BV) for 3.5-year-old volume (V_S,3.5) of the second-rotation were calculated with best linear unbiased prediction (BLUP) method and then used for selection of superior clones. Relative genetic gain (R_G) and selection efficiency (E) were assessed for the selected clones and the earlier first-rotation selections as compared to the V_S3.5-based-BV selection, respectively. 【Result】 The H² estimates for seven 3.5-year-old second rotation growth traits ranged from 0.29 to 0.48, indicating low to medium magnitude. The r_p and r_g among these second rotation growth traits were usually significant and positive. The r_p values between second rotation growth traits with 8-and 15-year-old growth of the first rotation were generally low and mostly not significant, while their r_g were generally moderate and statistically significant. A total of 41 clones were selected out from the V_S,3.5-derived-BV rank at an intensity of 15%, yielding in an R_G of 39.1%. As compared to the V_S,3.5-based-BV selection, the E values of BV selection in 8- and 15-year-old volumes (V₈ and V₁₅, respectively) of the first rotation were 2.8% and -2.0%, respectively, while the E values of multi-trait selection index (SI) method at ages 8 and 15 years of the first rotation were 6.1% and -1.6%, respectively, and those of multi-trait genotype-ideotype distance index (MGIDI) selection of the two first-rotation ages were 4.2% and 0.8%, respectively. 【Conclusion】 Growth cultivars at 3.5 years in the second-rotation of E. urophylla × E. tereticornis clones was under low to medium genetic control. Across-rotation growth traits have similar genetic basis. However, growth of the first rotation shows low predictability for that of the second rotation, and selections of the first rotation exhibit very low E values as compared to the V_S,3.5-based-BV selection, implying the necessity of across-rotation evaluation and selection at least in growth traits.