by species in 2-inch diameter classes. A subsample of 

 these understory trees by species and diameter class 

 was aged using increment borings and cross sections 

 at the ground line to determine what age classes were 

 present. Isolated understory trees were mapped by 

 X-Y coordinates. Groups of small understory trees 

 were mapped by recording the coordinates of several 

 points on the perimeter of the group. 



Ages of overstory trees were determined from incre- 

 ment cores taken 12 inches above the ground Une using 

 a power borer and bits up to 28 inches long (Scott and 

 Arno 1992). We bored trees repeatedly, if necessary, 

 to obtain a core that intersected or passed very close 

 to the pith. All overstory ponderosa pine were bored, 

 and in all plots most (74 to 95 percent) were sound 

 enough to obtain a good quality core. Many of the 

 Douglas-fir and western larch had advanced heart rot 

 and could not be aged; however, we were able to age 

 a large sample of both species in all size classes. Plot 

 B-4 contained four stumps of large ponderosa pine re- 

 moved in early 1900's logging. Three were from trees 

 >30 inches in diameter, a size-class represented by 

 only five living trees in the plot. Since these were 

 some of the oldest trees in the plot, we substituted for 

 the >30-inch logged trees by collecting ages from the 

 three nearest >30-inch pines outside the plot bound- 

 ary. These three tree ages were added to the tabular 

 stand age-class data, but not to plot tree-age maps. 



Cores were mounted into grooved boards in the field 

 using water-soluble glue, and each core was labeled on 

 the board (Arno and Sneck 1977). The boring height 

 above ground line and the direction of the core, in de- 

 grees clockwise from the uphill side, were recorded. 

 Generally, only the best core from each tree was kept 

 for analysis. 



In the laboratory, we used an orbital sander to pre- 

 pare the increment cores for measurement, first with 

 fine (150 grit) and then with very fine (400 grit) paper. 

 Annual rings were counted under a 7-30-power bin- 

 ocular microscope. Total tree age was estimated by 

 adding two correction factors to the raw count. The 

 first correction was the estimated number of years the 

 tree took to reach boring height, based on regeneration 

 data collected in past studies (Arno and others 1985; 

 Fiedler 1984). The second correction was the estimated 

 number of rings missed on an off-center core, computed 

 as the product of the estimated distance from the in- 

 nermost ring on the core to the pith and the growth 

 rate near the innermost ring. For example, 0.8 inch x 

 5 rings per inch = 4 years (Ghent 1955). Boring large 

 trees at 12 inches instead of the customary height of 

 4.5 ft, using long borers, and making repeated attempts 

 to intersect the pith, rather than extrapolating from 

 short or very off-center cores, improve estimates of 

 total age. Obtaining exact total ages requires destruc- 

 tive sampling— felling each tree and sectioning the 



stump at ground line. To assess accuracy of ages 

 determined from direct growth ring coxuits, increment 

 core ring-width measurements were dendrochrono- 

 logically cross-dated for the six plots measvu-ed during 

 the first year of the study (Fiedler and Steele 1992). 

 Cross-dating was conducted on 443 increment cores 

 using procedures developed by Holmes (1983, 1992), 

 which compare individual tree-ring series to a master 

 tree-ring chronology (the mean series obtained by aver- 

 aging all available increments for each year). 



Historic Age Structure and 

 Disturbances 



Age Determination Accuracy 



Results of the cross-dating procedure (Fiedler and 

 Steele 1992) indicate that 67 percent of the innermost 

 tree ring data determined from direct growth-ring 

 counts would be within 2 years of the actual age and 

 the majority of the remaining ring data would be 

 within 3 to 10 years of the actueJ age. Considering 

 the small errors associated with determining total 

 age to the pith at ground line, we felt these estimates 

 of total tree age were sufficiently accurate for charac- 

 terizing stand age structure by 20-year intervals 

 (table 1). 



Dry Sites 



Five of the six dry-site plots exhibited a nearly all- 

 aged structure (table la; figs. 3a,b). The remaining 

 plot (L-3) was dominated by a single broad age class 

 with only a small representation of other ages. The 

 data for plots L-1 and L-2 represent the nearly all- 

 aged structure of ponderosa pine and Douglas-fir 

 prior to 1900. Nevertheless, nearly adjacent plots 

 (L-1 and 2; B-1, 2, and 3) with virtually identical fire 

 histories varied substantially from each other in terms 

 of which groups of ages were present or most abun- 

 dant (table la). This suggests that despite similar 

 climate (and probably seed crops) in adjacent stands 

 on similar sites, spatial variation in tree establish- 

 ment, fire-caused mortality, and other factors perpetu- 

 ated a structural mosaic. 



In the nearly all-aged stands, despite fires at the 

 rate of two to four per century, ponderosa pine trees 

 that eventually ascended into the overstory became 

 estabhshed in 60 to 90 percent of the 20-year periods 

 between 1500 and 1900 (table la). Unknown addi- 

 tional age classes presumably were killed by fires. 

 The abundance of regeneration that developed into 

 overstory trees, termed "successful establishment," 

 varied through the decades, with an occasional major 

 pulse, such as between about 1525 and 1550, in B-2, 

 L-1, and L-2 (figs. 3a,b). Generally, however, pulses 



5 



