Family-specific responses in survivorship and phenotypic traits to different light environments in a seedling population of Fagus crenata in a cool-temperate forest

Abstract

Natural plant populations consist of individuals that exhibit variation in their phenotypic traits and demographic parameters. Here we report a study on maternal effects and the effects of different light environments on intra-specific variation in survivorship and ecologically relevant phenotypic traits of Fagus crenata seedlings in a cool-temperate forest community. We collected 901 seedlings from the ground beneath five maternal trees and used microsatellite DNA markers to identify maternal siblings that germinated naturally in the forest community. Selected seedlings were planted at three sites––one under a closed canopy with low light availability, one under a canopy gap in the natural forest community with moderate light availability, and one in a common garden with high light availability. The proportion of seedlings that were correctly assigned to their putative mother ranged from 60.0 to 82.7 % per maternal family, and 655 (72.7 %) seedlings in total were used for the analysis of survivorship and phenotypic traits. Among-family differences in survivorship remained after correcting for the effects of initial stem size during the first year after planting. However, this difference in survivorship became less pronounced in subsequent years. Seedlings grown under the canopy gap and/or in the common garden exhibited better performance in terms of phenotypic traits such as stem, leaf, and root morphology. In addition, seedlings of different maternal origins grown in the same environments had different individual leaf areas. These findings suggest that phenotypic variation due to maternal effects was a significant source of intra-specific variation within the local population.

Appendices

Appendix 1

The light environments at the three experimental sites were characterized by taking hemispherical photographs at 1.3 m above ground level near each block with an 8-mm Nikon fisheye lens in September and October of 2006 for the closure and gap site, respectively, and in July 2007 for the common garden. The photosynthetic photon flux density (PPFD) of diffuse and direct light above the overstory canopy surface and at 1.3 m above ground level in each block were estimated using the image analysis program, LIA32 (Yamamoto 2000). The Total Site Factor (TSF, ter Steege 1994) at the canopy surface of the seedling populations during the growing season was calculated using the PPFD values recorded directly above the overstory canopy surface and at 1.3 m above ground level in each block.

Appendix 2

In the permanent plot, we sampled single leaves from 74 and 51 individual seedlings growing under closed canopies and canopy gaps in 2006, respectively, and 30 and 22 in 2007. The individual leaf area was estimated using the following formula:

$$A = a \times (l_{1} \times l_{2} )^{b} ,$$

where A represents the estimated leaf area obtained using the image analysis program LIA32 (Yamamoto 2000), l ₁ is the leaf length (cm), l ₂ is the leaf width (cm), and a and b are coefficients. By fitting this expression to a linear regression model, it was possible to estimate the coefficients (Table 4), which were used to calculate the individual leaf area for each seedling. The values for 2007 were corrected using the proportion of healthy leaf area (=1 − proportion of leaf area damaged; see Fig. 2 in the main text).

Table 4 Coefficients and adjusted R ² values for the regression of individual leaf area against the product of leaf length and width in F. crenata seedlings

Appendix 3

See Table 5.

Table 5 Parameters of the hierarchical Bayesian models

Appendix 4

See Table 6.

Table 6 List of error distributions, link functions, and priors for the β _intercept in the hierarchical Bayesian models used in the analysis of variation in phenotypic traits