Skip to main content
. 2013 Mar 5;3(4):961–970. doi: 10.1002/ece3.518

Table 2.

Selection gradients (β) from multiple linear regressions for selection exerted on black pine by crossbills at two sites, Samarina and Aetia (n = 78 and 96 trees, respectively)

Samarina Aetia


Trait β ± SE 95% CI β ± SE 95% CI
Number of cones −0.0089 ± 0.0299 −0.0760, 0.0437 0.0032 ± 0.0019 −0.0076, −0.0005
DBH −0.0162 ± 0.0355 −0.0856, 0.0533 0.0039 ± 0.0017 0.0014, 0.0079
Cone width/length −0.0098 ± 0.0216 −0.0537, 0.0312 −0.0016 ± 0.0008 −0.0033, 0.0001
Scale thickness 0.0599 ± 0.0186 0.0247, 0.0985 0.0003 ± 0.0011 −0.0019, 0.0025
Number of full seeds −0.0149 ± 0.0177 −0.0486, 0.0205 0.0018 ± 0.0010 −0.0001, 0.0039
Number of empty seeds 0.0339 ± 0.0166 0.0026, 0.0705 0.0023 ± 0.0011 0.0006, 0.0048
Individual seed mass 0.0599 ± 0.0233 0.0146, 0.1075 0.0020 ± 0.0013 −0.0002, 0.0049

The overall models for Samarina and Aetia were significant (P < 0.0001 and P = 0.001, respectively). Selection gradients that do not overlap with the 95% confidence intervals (P < 0.05) are in bold. VIF < 2 for all traits in both models, |r| < 0.38 for all trait correlations.