a1 Nicholas School of the Environment and Center for Tropical Conservation, Duke University, P.O. Box 90381, Durham, NC 27708, USA
a2 Yale School of Forestry and Environmental Studies, Yale University, New Haven, CT 06511, USA
a3 Institute of Statistics and Decision Sciences, Duke University, P.O. Box 90251, Durham, NC 27708, USA
a4 Duke University Center for Tropical Conservation, 3705-C Erwin Road, Durham, NC 27705, USA
Inability to age tropical trees has imposed major limitations on the basic and applied science of tropical forests. Here advantage was taken of even-aged stands present in successional chronosequences found on Amazonian Whitewater river meanders to simplify the assumptions needed to estimate tree ages from growth measurements. Growth increments of eight common early successional species were measured in 21 0.5-ha plots evenly distributed over chronosequences from the earliest post-pioneer stage to mature Ficus-Cedrela stands representing approximately the mid-point of primary succession. Increment measurements, based on 4 or 5 y of growth, were arrayed in scatter diagrams against the midpoints of the growth intervals. A loess regression of the points, weighted for the higher mortality of slow-growing individuals, was then conducted to generate a ‘best estimate lifetime growth trajectory’ (BELGT) of a ‘typical’ individual surviving to maturity. The BELGT curves were integrated to generate a set of derived curves describing the time required by a ‘typical’ surviving individual to attain any given size up to the maximum for the species. Predictions of the ages of particular stands were derived from these latter curves and found to agree within 3 to 20% of ages independently estimated from the rate of point bar accretion.
(Accepted May 27 1997)
1 Author to whom correspondence should be addressed.