In an extensive series of trials with open-pollinated families of loblolly pine (Pinus taeda L.), resistance to fusiform rust disease (caused by Cronartium quercuum (Berk.) Miyabe ex Shirai f. sp. fusiforme) at individual test sites was relatively unpredictable for the families deemed most resistant. The most resistant families were also the most stable for performance across test sites, with stability defined as the slope of the regression of family means for rust infection versus site means for rust infection. A family's R-50 value (its predicted rust infection level when the site mean infection is 50%) was correlated to its stability parameter or slope (r = 0.78). On average, any one family's level of infection (% galled) was reasonably predictable for any given infection level at a given site; the average coefficient of determination (r2) was 0.78 for the regression of family means for rust infection versus site means for rust infection. However, the six most stable families for resistance had the lowest r2 values (average r2 = 0.58). We speculated that the lower predictability for the most resistant families was due to interactions of specific resistance genes in these families and corresponding avirulence and (or) virulence levels in the pathogen populations that may differ among sites. Although the predictability of the individual resistant families was relatively low, if these families were bulked into a resistant seed lot, they performed in a more predictable manner with r2 = 0.74 for the regression of the bulk mean versus site means. Bulks of four to six highly resistant families appeared to be a good solution to obtain stable and predictable performance across a range of sites.
Correspondence of Loblolly Pine Response for Fusiform Rust Disease from Local and Wide-Ranging Tests in the Southern United States
