Each forest range was treated as a sampling block, and each
management type (betel agroforest, teak plantation and unmanaged
secondary forest) was considered a treatment that occurred
in each forest range. For each treatment, we established six replicate
sites that were located across the forest ranges in an unbalanced
design, such that each range contained at least one
replicate of each treatment (see Appendix A). In each replicate site
we randomly located three forest plots for data collection. Therefore
we established 18 0.01-ha forest plots per treatment and a
total of 54 plots.
All stands used were similar in age class, prior management history,
distance to edges with other cover types, and previously
unburned according to detailed interviews with local residents
and Forest Department staff. The forests were state reserves from
the 19th century through the post-colonial period, comprising
mixtures of native forest and plantations of varying rotation
lengths. They were cleared from 1950 to 1966, then either abandoned
(becoming secondary forests) or planted with timber (teak)
or betel (agroforests). Details on study system and sampling are in
Appendix A.