Should more individuals be sampled when measuring functional traits of tree species in habitat-heterogeneous karst forests?

When measuring plant functional traits across geomorphologies, 5-10 healthy individuals of a plant species are commonly sampled. However, whether more individuals should be sampled in habitat-heterogeneous karst vegetation remains unknown. In this study, two dominant tree species (Clausena dunniana and Platycarya strobilacea) in karst evergreen and broadleaved mixed forests in Southwestern China were selected. On the basis of a large quantity of individuals of the two species grown in different peak clumps and slope positions, variations of 10 morphological traits in the two species were statistically analyzed. The suggested sampling number of individuals, which could mostly represent the common trait characteristics, was further explored. All traits showed significant differences between the two species (p < 0.05). The traits of P. strobilacea displayed larger intraspecific variations than those of C. dunniana, except for twig dry matter content. The bark thickness (BT), leaf area (LA), and specific leaf area (SLA) of C. dunniana and the BT, SLA, LA, leaf tissue density, and bark tissue density of P. strobilacea presented large intraspecific variations. Most traits exhibited significant differences between peak clumps and/or among slope positions (p < 0.05). Random sampling analysis indicated that the suggested sampling numbers of individuals for the 10 traits are 6-23 in C. dunniana and 9-29 in P. strobilacea. The common accepted sample size in normal geomorphologies is not sufficiently large in most cases. Larger sample sizes are recommended for traits, such as SLA, BT, and LA, with larger intraspecific variations. Therefore, under sufficient labor, material, and time, more individuals should be sampled when measuring plant functional traits in habitat-heterogeneous karst vegetation.