Of the 210 inoculated plants 86.67% showed disease symptoms. No plants from the host line Fin showed disease symptoms when inoculated with Emco spores. One test plant each in ten of the 216 transmission tests using uninoculated control plants became infected eight days later. These 4.63% of cases represent either contaminations between test plants in the same container or errors in notation. This gives us an average total transmission of 0.46 target leaves for all controls plants, which can be considered as the error of measure for this variable.
Transmission and Infection intensity
Transmission, estimated as the asymptote of the sigmoid curve fitted to the cumulated daily transmission data of the eight transmission events, differed among the different origins of the parasite strains, with the Orsay strains being the least transmitted, and among the six host lines. Transmission was highest on the host line Gb, followed by an indistinguishable group formed by Pyr, Tch and Tsu, followed by Sue and lowest transmission was on Fin (see Figure 1a). The interaction between host lines and origin of parasite strains was highly significant (Table 1). Indeed on the host lines Gb and Pyr the transmission was similar among the three origins whereas on Sue and Tsu the laboratory strains transmitted better than did wild strains. Furthermore, on the lines Pyr and Tch the strains from Fribourg transmitted more than those from Orsay (Figure 1a).
Transmission varied significantly by parasite strain, host line and their interaction even within some origins (Tables 1 and 2). Host line effect was always significant, the two parasite strains from Orsay differed significantly from each other, and significant interactions were found for Orsay and laboratory strains (Figure 1b and 1c).
The results were all qualitatively similar for the infection intensity on inoculated plants, measured as the number of infected leaves (Tables 1 and 2). All main and interaction effects remained significant when the host Fin, which was fully resistant to the parasite strain Emco, was excluded from the analyses.
Relation between host and parasite fitness
Different host lines had different fecundities across the 48 combinations, including controls. We found no general genetic correlation between host fitness (seed production) and parasite fitness (transmission) but the interaction between host line and this covariable was significant, revealing that the slope of this relationship was not homogeneous (Table 3). Rank correlations calculated separately for each host line across the eight parasite treatments (seven parasite strains and one control) detected a significantly negative genetic correlation only for the host line Gb (Spearman's rho = -0.81; P-value = 0.015) (Figure 2).