Parameter estimates and hypothesis testing for data from a retention period assay using a model of insect vector infectiousness (Model 1, S1 Appendix, see also Fig 1 for assay schematic).

Note that the model of insect vector infectiousness involved clearance of the pathogen by the insect and constant infectiousness up to clearance. In a), mortality probabilities were estimated from combining summary data representing cohorts that remain extant (see Table 1 row 1 for data) with Model 1. In b) probabilities of loss of infected insect and vector efficiency were estimated from combining summary data representing test plant infection (Table 1 row 2) with Model 1. Further statistical analysis tested the null hypothesis that there was no loss of infectious insect vectors beyond that predicted by insect mortality in the laboratory. The hypothesis involved the test statistic in c), which was derived from the posterior distributions in a) and b) (see Methods section in main text for further description). Posterior distributions were calculated using rStan v2.21.0 [33], and Model 1 (S1 Appendix) and posterior distributions are summarised throughout a − c by 50th (median), 2.5th and 97.5th percentiles. Note that here, and in all subsequent results tables, for simplicity figures are truncated to the third decimal point. All analyses were carried out in R version 3.63 [34]. Note that low insect survival in the experiment (c iii) was associated with a test statistic that did not significantly deviate from 1 (c i).