Statistical analyses
Fecundity was calculated as the sum of eggs laid within each mesh cage
by the 25 females released on the three knotweed plants within each of
the six replicated cages. Fecundity data were square root transformed to
meet the assumptions of normal distribution. An Analysis of Variance
(ANOVA) model was used with psyllid populations (Kyushu, FemKYU, FemHOK,
and Hokkaido) as a fixed effect to compare fecundity of the different
populations.
To evaluate oviposition preference the number of eggs laid on each plant
species by each psyllid population were compared using a linear mixed
model (lmer ) where psyllid population, knotweed species
(Japanese, Bohemian, and giant) and their interactions were the fixed
effects. Cage was included as a random effect to account for possible
differences among plants used in the different replicates. Egg count
data were square root transformed to improve normality.
A general linear mixed model (glmer ) was used to compare
development success of the different psyllid populations on the three
knotweed species. Given that the survival data is bounded between 0 and
1, a binomial distribution was used with a logit link function. Psyllid
population, knotweed species and their interaction were the fixed
effects and cage was the random effect in the model. Wald chi-square
test was used to obtain test statistics for this model.
For the oviposition preference and survival analyses post-hoc pairwise
comparisons were performed using the emmeans package with
Tukey-adjustment (Lenth, 2021). All analyses were conducted in R v.4.2.3
using the lme4 package (R core team 2022).