In May the seedlings were moved into a mist chamber 

 in preparation for inoculation. The seedlings were just 

 beginning their second growing season, and needles were 

 just beginning to break out of the fascicle sheaths. Mist 

 nozzles were operated to provide a constant visible fog 

 throughout the chamber. Twenty-four hours later, the 

 stems of the seedlings were wet and allowed to sit for 

 1 hour to let excess moisture drain. The seedlings were 

 then inoculated. 



Fresh inoculum was collected from at least 50 galls 

 from a young ponderosa pine stand located at the Lone 

 Mountain Tree Improvement site (lat. 47°54' N., 

 116°49' W., elevation 2,488 ft) about 25 miles north of 

 Coeur d'Alene, ID. Inoculation was accomplished on 

 May 30, 1984, by blowing spores over the seedlings using 

 an air-sprayer adapted from an Erlenmeyer flask. Each 

 replication was inoculated separately with a mixture of 

 1.2g of aeciospores and 7g of talc. This gave a spore-to- 

 seedling ratio of about 20,000 to 1. The seedlings were 

 kept in the mist chamber for an additional 72 hours mak- 

 ing sure that a visible fog was maintained. The mist 

 nozzles were then turned off and the seedlings allowed 

 to slowly dry. When dry, they were moved to a shade- 

 house. After a couple of weeks of acclimation, they were 

 moved to a nursery at Priest River Experimental Forest 

 (lat. 48°21' N., long. 116°52' W., elevation 2,400 ft), where 

 they were planted at a 6- by 6-inch spacing in beds 30 ft 

 wide. The 10 seedlings for each provenance in each repli- 

 cation were planted in two adjacent rows of five trees 

 each. Planting dates were June 19 and 20, 1984. 



Inspections for symptoms of infection were made for 

 each seedling at 3 months, 15 months, 27 months, and 

 39 months after inoculation. Data were tallied for the 

 presence of galls and reactions on the stem indicative of 

 defense reactions. The following variables were used to 

 assess variation among populations: 



1. Infected. Trees with galls at 39 months. 



2. Resistance. Trees with no galls at 39 months. 

 These were placed into three categories of resistance: 



a. No reactions. Trees that did not have galls or 

 bark reactions. 



b. Bark reactions. Trees that did not have galls 

 but had visible reactions on the stem indicative of defense 

 reactions. 



c. Gall death. Trees that had galls at 3, 15, or 27 

 months after inoculation for which the gall was dead at 

 39 months. 



3. Rate of gall appearance. Percentage of trees with 

 galls at 39 months that were visible at 3 months. 



4. Rust-killed trees. Trees with galls that died by 

 39 months. 



Analysis of variance was used for each variable to 

 determine significant differences among stands. Simple 

 correlations were performed for the rust variables with 

 3-year-old height, needle length, percentage dead due to 

 drought, and initiation of bud growth in a greenhouse. 

 The growth and drought data were from Rehfeldt (1986) 

 from different seedlings but from the same populations 

 and same seed collection. 



Simple and multiple regression models were used to 

 relate the rust variables to elevation and geographic loca- 

 tion of the seed source. The independent variables were 

 elevation, latitude, longitude, northwest departure, south- 

 west departure, azimuth from the inoculum source, and 

 distance from the inoculum source. Northwest (latitude 

 x longitude) and southwest (1/latitude x longitude) depar- 

 tures were derived by rotation of the grid of latitude and 

 longitude by 45 degrees. Multiple regression was used 

 when simple regression analysis of the rust variables 

 were significant. Squares of the above independent vari- 

 ables were added to accommodate the possibility of non- 

 linear patterns of variation. Therefore, 14 independent 

 variables were included in a stepwise regression for 

 maximizing R 2 (SAS 1982). In addition, the geographic 

 variables were nested within four geographic regions: 

 northeastern Washington, most of Idaho, northwestern 

 Montana including the most northerly stands in Idaho, 

 and middle to southwestern Montana (fig. 1). 



RESULTS 



The analyses detected differences among populations 

 for the degree of infection and in the kinds of resistance 

 reactions (table 1). Most rust factors varied randomly 

 with the environment; the degree of infection was the only 

 factor that was related to elevation or geographic origin of 

 the seed. 



The average level of trees with galls 39 months after 

 inoculation was 28 percent; populations varied from 3 to 

 63 percent (table 1 and fig. 2). Three major categories of 

 resistance were evident: (1) trees that expressed no 

 symptoms — 38 percent; (2) trees that expressed bark 

 reactions — 52 percent; and (3) trees in which the galls 

 died — 10 percent. Differences among populations for rate 

 of gall appearance and for death due to rust were high but 

 not significant. 



There were no strong associations between the rust 

 factors and the environmental factors (table 2). The R 2 

 for level of susceptibility and elevation was 0.15, the high- 

 est single association. Using all environmental factors 

 with the level of infection, the R 2 for the best fit stepwise 

 multiple regression model was 0.26. Figure 3 shows geo- 

 graphic lines at a constant elevation (average elevation 

 for all stands). The distance between contour lines equals 

 V2|7s<f(0.2)]. Thus distance equivalent to two contour 

 lines differs by a probability of about 0.2. When compar- 

 ing populations from the same elevation, populations from 

 the west central area — the area near the origin of the rust 

 inoculum — displayed the highest level of infection. From 

 here, the level of infection decreased in all directions. 



Figure 4 shows the association between elevation of the 

 seed source and the level of infection. The elevation clines 

 are keyed to the four geographic areas shown to figure 1. 

 And they all indicate that the degree of infection increases 

 as elevation of seed source increases, although the rela- 

 tionship tails off at high elevation. 



2 



