Figure 4— Geographic patterns of variation for 

 mesic fieigfit presented as isoplettis of equal 

 performance for base elevations, the lowest 

 elevation at v/hich ponderosa pine occurs in 

 a given area. The interval between isopleths 

 equals V2[lsd(0.2)]. Isopleths represent 

 positive or negative deviations from the mean 

 (x) of all base populations. 



about the 80 percent level of probability. In figures 3 and 

 4, the distance between isopleths (contour lines) is scaled 

 to V2[Zsd(0.2)]. This means that populations separated by a 

 distance equivalent to two isopleths are assumed to differ 

 with a probability of about 0.2. 



The relationship between the elevation of the seed 

 source and population performance is presented in figure 2 

 for four variables that span the range of responses for all 

 variables. Not shown are the elevational clines for ad- 

 justed mesic height and xeric height reduction, which 

 nearly duplicated that for mesic height. Clines for the 

 cessation and amount of shoot elongation were similar to 

 that of the duration of elongation, and clines for adjusted 

 xeric height, leaf length, and the initiation of shoot elonga- 

 tion failed to depict differentiation in excess of lsd{0.2). 



These elevational clines indicate that growth potential 

 decreases as elevation of the seed source increases. In con- 

 trast to populations from high elevations, those from low 

 elevations tend to be tall, they exhibit a late cessation, 

 long duration, and rapid rate of elongation, and under 

 xeric culture, their growth is reduced the most. 



The elevational interval associated with a mean differ- 

 ence between populations that equals lsdiO.2) is approx- 

 imately 440 m for the linear cline involving cessation of 

 shoot elongation (fig. 2). Although the interval for the 

 nonlinear clines would change as elevation changes (fig. 2), 

 population differentiation at the mean elevation (1,075 m) 

 can be detected across about 400 m for both mesic height 

 and adjusted xeric height, but about 450 m for rate of 

 elongation. 



Geographic patterns of variation at the mean elevation 

 are illustrated in figure 3 for the four variables that repre- 

 sent the range of patterns. Patterns for all variables not 



presented follow those for the duration of shoot elongation 

 and mesic height. Thus, when comparing populations from 

 the same elevation, populations from west central areas 

 display the largest values of traits reflecting growth 

 potential. From this area, growth potential declines in all 

 directions. 



Patterns for adjusted xeric height and leaf length de- 

 viate from the general. Figure 3 shows that if all popula- 

 tions had been the same height at age 2, those from the 

 dry lands to the east would have been tallest (largest ad- 

 justed xeric height) at age 3 on the xeric site. Geographic 

 clines for leaf length indicate that southern populations 

 have the longest leaves while those from the northwest 

 have the shortest. 



Because geography and elevation are not independent of 

 each other, genetic variation between populations at the 

 base elevation (fig. 4), the lowest elevation at which 

 ponderosa pine occurs in a given locality, is arranged 

 along much steeper clines than variation at a constant 

 elevation (fig. 3). As judged by the number of and distance 

 between isopleths, geographic chnes at the base elevation 

 are nearly twice as steep as those for a constant elevation. 

 Figure 4 shows that populations of the highest growth 

 potential should come from among base populations from 

 the west central region. From this area, growth potential 

 of base populations decreases regardless of whether eleva- 

 tion is decreasing toward the northwest or increasing 

 toward the south or east. The average distance between 

 differentiated populations occurs on a relatively small scale 

 (54 km) across the landscape. 



Similarities in patterns of variation between the vari- 

 ables (figs. 2 and 3) result from intercorrelations, some of 

 which are extremely strong (table 4). Populations that 

 were tall under mesic conditions exhibited other characters 

 associated with a high growth potential: a late cessation, 

 long duration, fast rate, and large amount of shoot elonga- 

 tion. Although these same populations tended to be the 

 tallest under xeric culture (table 4), they suffered the 

 greatest xeric height reduction in that the growth poten- 

 tial expressed under mesic culture was reduced the most 

 by xeric culture. 



DISCUSSION 



Results have demonstrated genetic differentiation be- 

 tween populations for a variety of loosely intercorrelated 

 traits. These intercorrelations, such as those in Pseudo- 

 tsuga menziesii (Rehfeldt 1983b), Pinus contorta (Rehfeldt 

 1983a), and P. ponderosa from central Idaho (Rehfeldt 

 1986), involve traits that are part of an annual sequence of 

 developmental events. This sequence begins with dehard- 

 ening in the spring, includes shoot elongation, leaf expan- 

 sion, bud development, and lignification, and concludes 

 with cold acclimation. The components are intercorrelated 

 because the entire sequence must fit into a growing 

 season of finite length. Thus, populations that inhabit en- 

 vironments where growing seasons are short cease devel- 

 opmental events early and are, therefore, short, even 

 when grown under optimal conditions. 



For ponderosa pine, a species that commonly occurs in 

 habitats that border the prairie, the growing season may 



6 



