500 



400 



300 



I '200 



100 



o 



■4 h 



o 



-cr 



o 



+ 



+ 



6 7 8 

 Day of Attack 



9 10 11 12 13 



-•- Baited v T1 



o T2 



Figure 2 — Cumulative number of attacks on the baited tree (Focus) and two 

 additional trees in plot that were successfully attacked. Day 1 is August 6, 1 995 

 (JD218). 



seventh tree attacked. Based on previous studies of 

 attack behavior, we assume that attack density on the 

 portion of bole sampled (0 to 2.4 m) is representative 

 of the whole tree. Rasmussen (1974) observed 79 per- 

 cent of initial attacks below 2. 1 m, with a mean attack 

 height of 1.4 m. Others observed that attacks are initi- 

 ated on the lower bole in a random fashion, with 

 attacks filling in the remainder of the bole with time 

 (McCambridge 1967; Reid 1963; Safranyik and 

 Vithayasai 1971). 



One day after removing the bait from the original 

 focus tree, beetle attacks occurred on four trees in the 

 plot; the closest attacked tree was a distance of 1.6 m 

 from the focus tree. At that time, the focus tree had a 

 greater number of attacks than any of the four sur- 

 rounding trees. Of four trees attacked, only two were 

 successfully mass attacked. A tree 0.58 m from the 

 focus tree was attacked on day 1, but because it was 

 attacked prior to removing the bait it was removed 

 from analysis. By day 3, a tree 7.1 m from the focus 

 tree had the greatest rate of attacks. Attacks on the 

 original, baited focus tree peaked on day 2, although it 

 continued to receive at least a few attacks through 

 day 7. Cumulative attacks on this tree did not level 

 out until after surrounding trees had been attacked 

 (fig. 2). This suggests that the switch in attacks to 

 surrounding trees occurred prior to full utilization of 

 the focus tree. 



If we only consider those trees that were consecu- 

 tively attacked from a nonbaited tree, a similar pat- 

 tern is observed (fig. 1). On the day of peak attack 

 density on T2 (day 3), new attacks occurred on T3. 

 The following day, at peak attack density of T3, T4 was 

 initially attacked (fig. 3). The next day, at peak attack 

 density on T4, T5a, and T5b were attacked (fig. 4). In 

 all cases, on the day when attacks switched to another 

 tree, the proportion of attacks on the current focus tree 

 was less than 45 percent (fig. 2, 3). These results again 

 suggest that attacks are redirected to a new tree prior 

 to full utilization of the tree currently under attack. 

 McCambridge (1967) observed that trees adjacent to 

 the focus tree came under attack by mountain pine 

 beetle before the initial focus tree had been fully mass 

 attacked. Anderbrandt and others (1988) observed a 

 similar pattern with Ips typographus Linnaeus. 

 Geiszler and others ( 1980), however, reported that the 

 switch to a new tree occurred after the original focus 

 tree had been fully mass attacked. In our plot, tree 

 spacing was closer than that reported in Geiszler and 

 others (1980). Closest distance between consecutively 

 attacked trees in our study was on day 5 between T4 

 and T5b, which were less than 1 m apart (fig. 1) 

 Geiszler and others (1980) reported the closest recipi- 

 ent tree to be 3.6 m from the focus tree. Because it is 

 impossible from our study design to say which trees 

 were being switched from and to, we use the general 



3 



