Eucalyptus tereticornis Sm. is an important species for agroforestry and commercial plantations in India. Results of a combined provenance–progeny trial of E. tereticornis laid out in 2002 at Midnapore (West Bengal) located in eastern tropical region of India are reported and discussed. Twelve provenances representing 70 families from Australia and Papua New Guinea were evaluated up to the age of 3 years. As a local seed source open-pollinated seed collected from a land race, i.e., Mysore gum (commonly known as Eucalyptus hybrid) was used to serve as check material (control). Significant variation in plant height, clean stem height, girth at breast height (GBH) and field survival was observed due to provenances, as well as families within provenances. This offers an ample scope to a breeder for increasing the growth and productivity in E. tereticornis through selection of provenances and families having potential for higher productivity. Heritability (narrow sense) values were appreciable for growth traits. Within provenance individual tree heritability estimates for height, clean stem length, GBH and number of branches at age 3 years were 0.318, 0.215, 0.269 and 0.231, respectively (assuming a coefficient of relationship of 0.4 for open-pollinated families of E. tereticornis). Provenance of Walsh River, Queensland Australia performed best for plant height, clean stem height, GBH, number of branches and field survival. Apart from this, the provenance from Oro bay to Emo, Papua New Guinea and Burdekin river, Queensland also showed good performance. Provenance from Yurammie, SF from New South Wales performed poorest for growth and survival. Correlations on growth traits were high and statistically significant, indicating that substantial gains could be achieved through indirect selection for one trait based on the direct selection for another. Age–age genetic correlations between age 1 and 3 years for the growth traits were highly significant and positive. Geographic clinal variation pattern was observed as latitude was negatively correlated with height and GBH ( P < 0.05); longitude with number of branches (negative, P < 0.05). Fair differences were observed between phenotypic and genotypic coefficient of variability.