And they are showing that there is no substitute for long-term data collecting to give real insights into forest growth. Douglas Sheil, a CIFOR researcher, revisited tree plots that were originally established during the 1930s and 1940s by W.J. Eggeling, an English forester, in Budongo, a Ugandan rain forest. The Budongo plots are believed to be the oldest surviving in the humid tropics.

"We need a better way to predict how tropical forests react to external influences," Sheil said. "In this case, the long-term plots allowed us to recognise potentially major, yet unanticipated, drivers of forest change." These influences leave their impression on the process of succession in the forest that can be seen in the analysis of age location and composition.

In Sheil's opinion, these plots offer a unique window on long-term change. "We have all learnt to place great importance on spatial sampling as this determines many of the choices and assessments we make as researchers," observed Sheil. "We have forgotten the worth of extended study periods to observe reality."

Old forest plots remain a rich vein of understanding and insight even though replication is a problem. "Surviving plots are often single," Sheil said, "But the continuous cycle of crosschecking and verification over time lets us identify and correct problems that are considerably more prevalent, though rarely acknowledged, in short-term studies."

"Careful observation and description provides the material for understanding," concluded Sheil. "Permanent plots should play a central role in our studies of tropical forest but we need to recognise both their benefits and limitations."

Sheil, D. 2003 Observations of long-term change in an African rain forest. In: Hans ter Steege (Ed.) Long-Term Changes in Tropical Tree Diversity as a Result of Natural and Man Made Disturbances: Studies from the Guiana Shield, Africa, Borneo and Melanesia, Tropenbos Series 22:37-59. Wageningen, the Netherlands.