of lodgepole pine and some ponderosa pine and larch 

 regenerated following underbums in about 1839, 

 1850, and 1861 (fig. 3c). Many of the smaller over- 

 story lodgepole pine and larch trees have miiltiple 

 scars from fires that occurred after their establish- 

 ment. Scattered individuals and small groups of 

 >400-year-old ponderosa pine and larch survived the 

 1817 fire and are found throughout this stand. 



Supplementary plot F-3 was also in a young even- 

 aged serai ponderosa pine stand that established fol- 

 lowing a replacement fire (about 1861), additional 

 scattered survivors had become established around 

 1660 (fig. 4). Supplementary plot F-4 was located in a 

 mosaic of two even-aged classes: ponderosa pine dating 

 from about 1600; and lodgepole pine/ponderosa pine/ 

 larch dating from the late 1800's. All four Flathead 

 plots had evidence of fires at average intervals of 20 

 to 30 years prior to 1900; the majority were nonlethal 

 surface fires. However, replacement burning also oc- 

 curred at intervals between 150 and 400+ years (fig. 4), 

 often in a patchy or mosaic pattern on the landscape. 



The higher productivity of the moist site-types 

 (Pfister and others 1977) is reflected in greater under- 

 story fuels. For example, the Flathead National For- 

 est stands have duff mounds 16 to 24 inches thick at 

 the base of large pines, while dry-site stands have 

 much smaller accumulations (2 to 6 inches). Duff 

 mounds on the Bitterroot moist site (B-4) were of inter- 

 mediate depth. It is unlikely that stand-replacement 

 fire would have occurred in the relatively open pre- 

 1900 stands on the moist sites (Ayres 1901) unless 

 preceded by a buildup of fuels (Anderson and Brown 

 1988). Unusually long fire intervals could have facili- 

 tated replacement burning, but the fire history does 

 not suggest that the pi-e-1900 replacement fires were 

 linked to this. Another possibility is that bark beetle 

 epidemics caused extensive mortality and a buildup 

 of fuel from the dead trees that supported replace- 

 ment burning. Low vigor or stressed old growth pine 

 are especially vulnerable to the western pine beetle 

 (Dendroctonus brevicomis) (Johnson 1972). Mountain 

 pine beetle (D. ponderosae) epidemics have been com- 

 mon during this century in the extensive lodgepole 

 pine forests that surround our Flathead National 

 Forest plots, and these outbreaks can spread to pon- 

 derosa pine (Gibson 1993). 



Spatial Relationships in Old 

 Growth 



Photographs of pre- 1900 ponderosa pine stands 

 in the Northern Rockies (Ayres 1901; Leiberg 1899; 

 Wickman 1992) show a uniform-appearing overstory 

 and scarcely any understory. These stands had a 

 continuous fiielbed of highly combustible pine-needle 

 litter and dry grass, which allowed fires to spread 



fi"eely over large areas. Nevertheless, adjacent stands 

 (L-1, 2, and B-1, 2, 3) had noticeable differences in tree 

 composition and age structure. Plots L-l and L-2 were 

 only 200 ft apart on the same smooth slope, and they 

 had the same fire history between 1633 and 1919. Note, 

 however, their differences in circa- 1900 and modem 

 stand structure (tables la, 2; figs. 5a,b). For example, 

 in L-l only about 25 percent of the overstory trees 

 were Douglas-fir, compared with over 60 percent in 

 L-2 in 1900, prior to an episode of heavy mortality. 



In pre-1900 stands on most of the dry sites and on 

 some moist sites (B-4) the effects of individual fires on 

 mortality, regeneration, and establishment resulted 

 in a fine-grained, subtle mosaic of overstory trees of 

 various ages. This fine mosaic is identifiable in our 

 tree-age plot maps (figs. 5a-i), but is difficult to dis- 

 tinguish on the ground. For instance, note the six 

 290-year-old pines grouped in the lower left portion of 

 plot B-1 (fig. 5d), the 300-year-old pines and firs in the 

 upper center of B-3 (fig. 5f), and the three 530-year-old 

 pines in the upper right of L-3 (fig. 5c). 



In addition to the fine mosaic, occasionally on dry 

 sites we could readUy observe a coarser mosaic (units 

 of V2 to 3 acres) apparently related to patches of over- 

 story mortality caused by fire and other factors. For 

 example, note the patch of about 90-year-old trees in 

 the lower left of plot L-l (fig. 5a) that arose where the 

 1889 fire had caused mortality. In contrast, on the 

 Flathead moist sites the dominant spatial pattern 

 was a coarse mosaic of even-aged pine stands linked 

 to the patchy stand-replacement fires. 



Understory Development 



On both dry and moist sites, the major changes in 

 stand structure between 1900 and the 1990's were 

 an increase in basal area and in number of trees per 

 acre as well as the development of an understory of 

 shade-tolerant trees (tables 2 and 3). Douglas-fir 

 represented the majority of the understory trees in 

 all plots (figs. 6a-e, 7). Seven of the nine plots, all 

 those without extensive overstory mortality, had 

 post-1900 increases in basal area of 23 to 144 per- 

 cent, with a mean of 76 percent. (We were unable 

 to account for small trees that may have died since 

 1900, but small trees have little effect on stand basal 

 area.) Increased basal area levels in these drought- 

 susceptible forests probably contribute to mortality 

 fi-om insect or disease epidemics, which are becoming 

 widespread in serai ponderosa pine forests (Mutch 

 and others 1993; Wickman 1992). Two of the dry-site 

 plots sustained a large amount of mortality in recent 

 decades (table 2), due in part to root disease (species 

 unidentified); their basal areas have remained simi- 

 lar (L-2) or declined substantially (B-3) since 1900. 

 The Bitterroot National Forest moist site (B-4) has 



10 



