(1986) models also did not adequately explain the vari- 

 ation in growth of red fir. His best equations, developed 

 for the 2-year period just after thinning, explained less 

 than two-thirds of the variation. Equations for 8 years 

 following treatment had CD's of only 0.33 for diameter 

 growth and 0.24 for height growth. Other CD's reported 

 for height growth were: 0.35 (Ferguson and Adams 1980); 

 0.11 to 0.58 (McCaughey and Schmidt 1982); 0.56 (Helms 

 and Standiford 1985). The highest CD's developed were 

 0.77 for subalpine fir and 0.72 for spruce height growth in 

 Alberta, Canada (Johnstone 1978). Models for diameter 

 growth generally had higher predictive power than those 

 for height, supporting our findings at Coram. Comparison 

 of CD's across all studies may not be entirely valid be- 

 cause the nature of the dependent variable differed 

 among studies. Nevertheless, the mediocre CD's simply 

 reflect the inherent difficulty in identifying and measur- 

 ing key factors affecting tree growth in the field. The 

 models reported in the literature are generally state of the 

 art and are likely about the best that can be expected. 



Mortality and serious damage caused by logging in our 

 study area were reported earlier (Benson and Gonsior 

 1981) so we did not include this type of damage in the 

 present study, except that we included mortality caused 

 by logging as part of the basal area change. They re- 

 ported that more than 40 percent of the trees were either 

 killed or seriously damaged during the harvesting process 

 in the treatments where all the overstory was removed 

 and about 30 percent where only part of the overstory was 

 removed. We documented only mortality that was not 

 obviously caused at the time of harvest. Mortality of 

 advance regeneration less than or equal to 7 inches d.b.h. 

 was least in the complete overstory removal treatments 

 and controls (13 percent) and greatest in the partial re- 

 moval (25 percent), amounting to about 1 to 2 percent per 

 year, respectively. Among species, a higher proportion of 

 Douglas-fir (22 percent) died than of Engelmann spruce 

 and subalpine fir (13 percent). We do not know why mor- 

 tality was greater in the partial removals nor why a 

 higher proportion of Douglas-fir died than the other spe- 

 cies. Perhaps root diseases were more active, but this 

 presumption is highly speculative. About all we can say 

 is that our mortality rates are, in a general sense, similar 

 to those predicted by Hamilton (1986) for stands in north- 

 ern Idaho. He predicted 10-year mortality rates of 0.13 in 

 thinned and nonthinned stands on productive sites in 

 northern Idaho, which translates to 1.3 percent per year. 



A serious consequence of leaving advance conifer regen- 

 eration in the Northern Rocky Mountains is an increase 

 in stand susceptibility to western spruce budworm. West- 

 em larch and Douglas-fir when serai tend not to support 

 populations of the insect (Carlson and others 1985). 

 Stands dominated by subalpine fir, Engelmann spruce, 

 and/or grand fir in the overstory or understory are highly 

 susceptible. Two to three years prior to and during estab- 

 lishment of this study in 1973, budworm was active and 

 moderate in the study area; 30 to 50 percent defoliation 

 was present on the advance regeneration. Subalpine fir 

 and Engelmann spruce sustained more defoliation than 

 Douglas-fir. This supports other observations we have 

 made; when Douglas-fir is serai it tends to have less defo- 

 liation than its more shade-tolerant associates (Carlson 



and others 1985). We thought this situation would pro- 

 vide an excellent opportunity to assess differential effects 

 of the insect on advance regeneration. But for unknown 

 reasons, budworm nearly disappeared from the study area 

 after 1975, and we ended up with only 1 year's defoliation 

 data. Even so, we did find a significant effect of defolia- 

 tion on one species. Increasing defoliation was associated 

 with decreasing height growth of Douglas-fir (fig. 2). 



Other studies concerning the influence of defoliation on 

 tree growth support these results. Ferguson (1988) dem- 

 onstrated that probability of top dieback increased with 

 increasing defoliation, and that periodic height increment 

 was reduced. His study specifically considered defoliation 

 over a 5-year period — longer than our outbreak. Further- 

 more, at the time he selected his stands, the budworm 

 outbreak was in progress, therefore his sample trees 

 likely were defoliated for more than 5 consecutive years. 

 Proportion of retained foliar biomass was positively re- 

 lated to growth of Douglas-fir and grand fir in central 

 Washington (Nichols 1988), implicating the negative ef- 

 fects of defoliation. In our study we were not able to dem- 

 onstrate an effect of defoliation on height growth of sub- 

 alpine fir or Engelmann spruce, even though they sus- 

 tained more defoliation than Douglas-fir. Perhaps these 

 species, being more shade-tolerant than Douglas-fir, were 

 better able to recover following insect feeding, masking 

 the effect of only 2 years of defoliation. 



The amount of defoliation we observed was strongly 

 influenced by tree stature; other variables were of lesser 

 importance. Taller trees of all species incurred more 

 defoliation than smaller trees (table 7). This was ex- 

 pected because budworm tend to feed near the tops of 

 trees but drop down when foliage supplies become limit- 

 ing. In light to moderate outbreaks such as we experi- 

 enced in this study, foliage likely was not a limiting factor 

 for the insect so they probably did not move downward 

 very much. Furthermore, budworm larvae on small trees 

 less than 3 feet tall are easy prey for birds and ants — at 

 least half of the larvae on small trees may be removed 

 (Carlson and others 1984). Because defoliation increased 

 with increasing height for all species, potential crop 

 trees — the taller ones with more live crown — may be more 

 at risk than smaller trees. Douglas-fir trees with poor 

 crown ratios, in plots with lots of variation in height, 

 received the most defoliation. Subalpine fir trees with 

 poor prerelease 10-year radial growth, on plots with low 

 variation in height, were impacted most by budworm. 

 The common thread between defoliation on Douglas-fir 

 and subalpine fir appears to be some measure of tree 

 vigor. Poor crown ratios indicate poor vigor for Douglas- 

 fir, whereas poor prerelease growth rate seems to be the 

 indicator for subalpine fir. 



MANAGEMENT IMPLICATIONS 



This study suggests that culturing advance regenera- 

 tion on productive sites in the Northern Rocky Mountains 

 may not be wise if the primary objective is to produce 

 timber. Volume production may be mediocre compared to 

 yields expected from serai, even-aged regeneration estab- 

 lished following harvest. Furthermore, the increased sus- 

 ceptibility of advance regeneration to western spruce 



11 



