Systematic sampling of a large forested area, as done here, avoids the problem of subjectivity in selection of sample sites. For example, Munger’s (1912) principal objective was to provide information on potential future yields so he selected “well-stocked areas”; he acknowledges that his selected stands may be “high” in stocking and not representative of the average ALK tumor conditions due to the exclusion of areas of lower density and of the gaps and openings typical of dry forests (Munger 1912). Reference

data for small trees are rare; among the cited studies only Munger, 1912 and Munger, 1917 provides this information (Table 6). Few records exist and reconstructions are limited by availability of evidence (live and dead trees), since small trees are much more ephemeral than large trees – e.g., increasingly vulnerable to loss over time due to fire, insects, disease, and decomposition (Fulé et al., 1997, Harrod et al., 1999 and Mast et al., 1999). However, Moore et al. (2004) have demonstrated the potential for reasonable accuracy in reconstructing historical forest conditions. For central and south-central Oregon, Munger, 1912 and Munger, 1917 record of stand structure and composition for 93 ha of ponderosa pine-dominated stands in Klamath, Lake, and Crook counties was the only one that we could find for trees smaller than 50 cm dbh. Density of small trees

(15–53 cm dbh) LBH589 research buy was 8, 80, and 81 tph in Munger’s three samples; these records are well within the range (0–227, mean = 38, SD = 26 tph) recorded in our more spatially extensive and systematic sample. The singular exception to the congruence between our conclusions from the historical inventory and other existing historical records and reconstructions is a recent study (Baker, 2012) suggesting that approximately half the Chiloquin study area supported forests with a density of >143 tph. Baker (2012) reconstructed historical forest conditions in eastern Oregon using General Land Office (GLO) survey data, which consist of eight trees per section (64 ha). Four townships (T35-36S Thiamine-diphosphate kinase R8-9E) in his study area overlap our Chiloquin study

area. GLO survey data collected 1866–1895 would include a record of ∼1152 trees marking section and quarter section corners in this four township area while the BIA timber inventory includes 1,63,558 trees on 1355 transects. Density recorded in the BIA timber inventory across all habitat types ranged from 0 to 296 tph with a mean density of 60 ± 37 tph and a 95th percentile value of 132 tph for the same four township area. Reconstructed tree density based on GLO data (Baker, 2012) is nearly 2.5 times the mean tree density recorded in the timber inventory for the same area leading us to conclude that the Baker (2012) reconstruction significantly overestimates historical tree densities on the Reservation. We found that densities of 143 tph or greater occurred in fewer than 106 ha (3%) of the 3789 ha inventoried between 1914 and 1922 in the four township area.