cedure reduce the variance of the dependent variable (tree 

 mortality) more than any other pair of subgroups that could 

 be generated. Results are most readily visualized as a 

 branching diagram that details the mortality configuration 

 for individual trees in response to gypsy moth infestation. 



Results 



AID analysis of individual trees retained only three of the 

 many characteristics analyzed as independent variables in 

 its classification of tree mortality; tree crown condition, 

 crovi^n position, and species. A significant relationship exists 

 between each of these variables and tree mortality. 



Trees with poor crowns (more than 50 percent dead 

 branches, very thin foliage, or heavy sprouting) have low 

 vigor and are more likely to die after defoliation: 



Cumulative 

 mortality 

 1979-85 

 (Percent) 



Crown condition 



Good 

 Fair 

 Poor 



10 

 21 

 49 



Suppressed trees are often environmentally stressed and 

 less likely to survive the added stress of defoliation: 



Crown Position 



Dominant & co-dominant 



Intermediate 



Suppressed 



Cumulative 

 mortality 

 1979-85 

 (Percent) 



14 

 18 

 28 



Oaks, especially black and chestnut oak, are favored food of 

 the gypsy moth larvae and are usually attacked more 

 severely than other tree species, leading to higher death 

 rates: 



Species 



Black and chestnut oak 

 Other oaks 

 Other species 



Cumulative 

 mortality 

 1979-85 

 (Percent) 



25 

 19 

 14 



The AID analysis delineated ten final groups of trees and 

 the probabilities of mortality in each group (Fig. 1). Black 

 and chestnut oak trees with intermediate or suppressed 

 crown position, and poor crown condition (Group 1), had 

 the highest probability of mortality; 64 percent, as com- 

 pared to 19 percent for the total sample. This group com- 

 prised 4 percent of the total sample. Trees of other oak 

 species that are intermediate or suppressed and have poor 

 crowns (Group 2) also had high (54 percent) probabilities of 

 mortality. 



The combination of fair tree crown condition, intermediate 

 or suppressed crown position, and oak species gives a tree 

 a 35 percent probability of mortality (Group 5). 



Trees with good crowns have the lowest mortality prospect; 

 even black and chestnut oaks rated only a 1 in 8 chance of 

 dying if they had good crowns and dominant or co- 

 dominant crown position before the outbreak (Group 9). 



Tree mortality figures for all combinations of crown condi- 

 tion, crown position, and species group in the central Penn- 

 sylvania gypsy moth episode are shown in Table 1. Again, 

 the importance of these predictors— especially crown 

 condition— as key indicators of potential mortality is clear. 

 Regardless of species or crown position, more than one- 

 third of the trees with poor crowns died. In each species 

 group, vulnerability was also consistently greater for trees 

 with suppressed crown position versus intermediate or dom- 

 inant and co-dominant positions. More than one-third of the 

 suppressed black and chestnut oak trees died; even those 

 that had good crown condition prior to infestation. 



Table 1— Percentage of tree mortality associated with gypsy 

 moth in central Pennsylvania (1978-85), by initial crown 

 condition, crown position, and species group. 



Crown 



Crown 



Black and 



Other 



Other 



condition 



position 



chestnut oak 



oak 



species 









Percent - 















Poor 



Suppressed 



67 



62 



48 





Intermediate 



60 



41 



39 





Dominant and 











co-dominant 



35 



42 



40 



Fair 



Suppressed 



44 



46 



22 





Intermediate 



31 



30 



19 





Dominant and 











co-dominant 



17 



14 



14 



Good 



Suppressed 



34 



19 



13 





Intermediate 



21 



13 



6 





Dominant and 











co-dominant 



13 



8 



5 



2 



