41 



A closer inspection of Table 7 indicates that the five clusters 

 have modal peaks that correspond, with varying amplitude, to the five 

 hydrologic zones. Trees in cluster 1 are found almost exclusively in 

 Zones 2 and 3. Likewise, trees in cluster 2 occur most commonly in 

 Zone 3. Cluster 5 had the most restrictive distribution, with three of 

 the four species occurring exclusively in Zone 6. The remaining clus- 

 ters, 3 and 4, had less distinctive modes, but exhibited greater distri- 

 butions in Zones 4 and 6. 



Cluster analysis on the entire sapling and shrub data set did not 

 produce distinct groupings (Figure 6). Silver bell (HADI) was not eval- 

 uated further because it occurred at a single site. The data set was 

 split into the two sapling and shrub (bush) components, and cluster 

 analysis was recalculated on each component. The saplings alone 

 (n - 17) grouped more distinctly (Figure 7) than the shrubs (n = 11) 

 (Figure 8). Shrubs, therefore, were not used in any further analyses. 

 Saplings alone grouped into five distinct groups but not as strongly as 

 trees. The relative frequency of occurrence and cluster membership of 

 the sapling species are given in Table 8. 



The vine data produced five clusters, morning glory (IPWR) being a 

 single species cluster corresponding to Zone 2 (Figure 9) . Two vine 

 species poison ivy (TORA) and trumpet creeper (CARA) were not included 

 in larger clusters, largely because of their apparent wide tolerances of 

 hydrologic conditions. Table 9 depicts the relative abundance distribu- 

 tions of each vine species across the hydrologic gradient within a clus- 

 ter and illustrates a pattern similar to but not as strong as that of 

 the tree species. 



