Environmental contribution to needle variation among natural populations of Pinus tabuliformis.

Variations in the phenotypic characteristics of conifer needles is a consequence of genetic evolution that has been widely used in geographic variation and ecological studies. Although many studies are based on an in situ sampling strategy and generally realize the contribution of environmental effects to variation in needle traits, it is still uncertain which needle traits are most influenced by genetic effects and which are most influenced by the environment. Using both a common garden experiment to eliminate environmental heterogeneity and an in situ sampling strategy, we compared 18 Pinus tabuliformis needle traits among 10 geographical populations. Using both sampling strategies, we found significant differences in needle traits among populations and among individuals within populations. Differences in the "among-population" variance component between the two sampling strategies revealed the environmental contribution among natural populations for each trait. The among-population variance in the following traits exceeded 8%: needle length, number of stomata within 2 mm (NS2), number of stomatal lines on the planar side, number of resin canals (RCN) and the resin canal area (RCA). For the stability of needle traits, NS2, RCN, RCA, ratio of the vascular bundle area to the RCA (VBA/RCA), and MA/RCA differed significantly in more than five provenance changes between the common garden populations and natural populations, which may be susceptible to environmental effects. Conversely, the cross-sectional area, mesophyll area (MA), MA/(VBA + RCA), and MA/VBA were phenotypically stable. Geographic variation patterns and systematic relation of needle traits differed between the two sampling strategies, suggesting that in situ sampling results may reflect environmental effects and deviate statistical parameters for genetic study. Future studies of genetic evolution in the context of geographic variation should be based on appropriate sampling strategies and stable phenotypic traits. [ABSTRACT FROM AUTHOR]