38 



species accounted for 96.5 percent of all individuals and were the spe- 

 cies analyzed statistically. The herbaceous ground-cover data set con- 

 tained 268 species. Most occurred only rarely in this study, and only 

 30 species whose abundances equaled or exceeded 1 percent of the total 

 herb individuals were included in the analysis. These species accounted 

 for 48.9 percent of all individuals in this vegetative category. 



Cluster Analyses 



A method of calibration suggested by Ter Braak and Prentice (1988) 

 is cluster analysis in which an environmental value (hydrologic zone) is 

 predicted through use of species abundance indicators (relative fre- 

 quency) . In addition to relative frequency, cluster membership was used 

 as a predictor of hydrologic zone. Species were clustered according to 

 similar abundance distributions across the hydrologic gradient. 



Cluster analysis was used to group the 30 tree species into five 

 clusters based on the five hydrologic zones. Results (Figure 5) show 

 that, with the exception of chinaberry (MEAZ) and deciduous holly 

 (ILDE), five distinct groups can be discerned. Table 7 gives the rela- 

 tive frequencies of occurrence of each species in each hydrologic zone 

 (species are grouped by cluster) . Inspection of the data reveals why 

 chinaberry and deciduous holly did not group readily. Chinaberry is the 

 only species that occurs almost exclusively in Zone 5. Its occurrence 

 is also restricted to a single site. The distribution of deciduous 

 holly peaks in Zones 2 and 3 , but not to the extent of other common 

 species in these zones. Because chinaberry and deciduous holly did not 

 group readily, they were eliminated from further analysis. 



