Fig 1.
Relative ability of fungal isolates to kill H. armigera larvae depended on crop leaf diet.
Dots indicate mean mortality at day 14 post-infection; whiskers give 95% binomial confidence limits for each combination of pathogen treatment (on the x-axis) and plant leaf diet (in panels). Total n = 3811 larvae across the 9 treatment combinations.
Fig 2.
Diet and infection treatment strongly alter the relative fitness of different half-sibling families.
Each point and whisker represents mean mortality and 95% binomial CI for sire half-sibling families 14 days post-exposure, depending on plant diet (columns) and pathogen treatment (rows). Four of these families are highlighted consistently across panels (see coloured points) to illustrate contrasting patterns of performance across habitats. The sires are arranged along the x-axis in order according to their rank performance in each panel. The phenotypic patterns in this plot are formally decomposed into the additive genetic components in Table 1 and Fig 4.
Table 1.
Changes in plant diet are associated with strong genetic trade-offs. Below are summaries of genetic variances (orange, diagonal), covariances (blue, upper off-diagonal), and correlations (green, lower off-diagonal) for mortality across 9 combinations of plant and pathogen exposure treatments. Within-plant diet correlations are shaded in darker green to call attention to their consistently higher values than the cross- plant diet correlations.
Fig 3.
A substantial fraction of variation in larval survival in each treatment is heritable.
Posterior distributions from a Bayesian analysis of heritabilities (the proportion of variation in survival that is additive and genetic) in each of the nine combinations of plant diet and infection treatment. The density of each distribution indicates the weight of posterior evidence for a particular heritability value (on the x-axis). These estimates were obtained from 58 half-sib families of H. armigera fed on one of three leaf diets (soybean, maize, or tomato) and exposed to one of three pathogen-exposure treatments (M = M. anisopliae, B = B. bassiana, — = Control).
Fig 4.
Genetic correlations are lower when measured across environments that differ in plant diet.
Posterior distributions for cross-environment genetic correlations for mortality in H. armigera larvae grown in 9 different combinations of plant diet (soybean, maize, and tomato) and pathogen treatment (control, Beauveria, and Metarhizium). The 36 posteriors are clustered in the figure above depending on the axes of environmental difference, with environments differing in only 1 dimension on the left, and those differing in 2 dimensions on the right. Summaries of central tendency in the genetic correlations appear in the lower off-diagonal of Table 1.
Fig 5.
Schematic representation of the experimental design.
Second instar larvae from each female were randomly assigned to one of 9 treatments. Insects were exposed to one of three different infection treatments and were reared on one of three different plants. Larval survival was recorded daily thereafter.