Table 2.

Statistical summaries of the best models of juvenile growth rate for the one-month and six-month field studies. (The first row for each analysis contains p-values for each of the included predictors (values in bold are significant, α < 0.05). Predictors left blank were excluded from the final model. Partial η² quantifies variance explained by each predictor after controlling for all other predictors. Model β coefficients were normalized by the global standard deviation (1.679 × 10⁻³ or 1.195 × 10⁻³, respectively); resulting values indicate the effect (with 95% confidence intervals) of each term as a multiple of standard deviation, relative to the reference level. No equally parsimonious models were found; model weights were 1.00 and 0.76, respectively.)

one-month experiment
origina	treatmentb	origin × treatmentc	initial shell length
p = 0.992	p < 0.001 exp = 2.45 (2.12, 2.79) prot = 0.75 (0.41, 1.08)	p < 0.001 highprot × exp = −0.39 (−0.88, 0.09) lowexp × exp = −0.86 (−1.31, −0.40) lowprot × exp = −0.88 (−1.34, −0.42) highprot × prot = −0.03 (−0.51, 0.45) lowexp × prot = 0.27 (−0.18, 0.73) lowprot × prot = 0.26 (−0.20, 0.72)	p < 0.001 length = −0.10 (−0.13, −0.08)
six-month experiment
	treatmentb	time intervald	initial shell length	time × treatmente
	p < 0.001 prot = −1.53 (−2.06, −1.01)	p < 0.001 1–6 = −1.32 (−1.97, −0.67)	p = 0.006 length = −0.12 (−0.20, −0.03)	p < 0.001 prot × 1–6 = 1.57 (0.82, 2.33)

^aReference = high exposed.

^bReference = common garden for one-month experiment, exposed site for six-month experiment.

^cReference = high exposed, common garden.

^dReference = 0–1 month.

^eReference = 0–1 month × exposed.