Population diversification can be shaped by a combination of environmental factors as well as geographic isolation interacting with gene flow. We surveyed genetic variation of 243 samples from 12 populations of
Calocedrus formosana
using amplified fragment length polymorphism (AFLP) and scored a total of 437 AFLP fragments using 11 selective amplification primer pairs. The AFLP variation was used to assess the role of gene flow on the pattern of genetic diversity and to test environments in driving population adaptive evolution. This study found the relatively lower level of genetic diversity and the higher level of population differentiation in
C. formosana
compared with those estimated in previous studies of conifers including
Cunninghamia konishii
,
Keteleeria davidiana
var.
formosana
, and
Taiwania cryptomerioides
occurring in Taiwan. BAYESCAN detected 26
F
ST
outlier loci that were found to be associated strongly with various environmental variables using multiple univariate logistic regression, latent factor mixed model, and Bayesian logistic regression. We found several environmentally dependent adaptive loci with high frequencies in low- or high-elevation populations, suggesting their involvement in local adaptation. Ecological factors, including relative humidity and sunshine hours, that are generally not altitude related could have been the most important selective drivers for population divergent evolution in
C. formosana
. The present study provides fundamental information in relation to adaptive evolution and can be useful for assisted migration program of
C. formosana
in the future conservation of this species.