INTERPRETATION 



It is sometimes held that multivariate analysis offers an objective al- 

 ternative to traditional methods of classification. Mcintosh (1967) seems 

 to suggest for example, that ordination offers such an alternative. Lambert 

 and Dale (1964) however, note that subjectivity cannot be avoided in multi- 

 variate analysis in either sampling or interpreting results. 



In sampling, subjective choices must be made concerning which attributes 

 of the plants and sites will be measured and how, and which areas will be 

 sampled and how. In analysis there are many questions of technique, starting 

 with selecting a similarity measurement. Some of these questions have been 

 addressed in this appendix. 



At the completion of the multivariate analysis then, objectivity cannot 

 be claimed, but it produces one important advantage over some traditional 

 classifacatory techniques, which is that anyone using the same data and 

 techniques could produce identical results. It would then be possible to pick 

 one or more similarity levels and conclude that there are as many types as 

 there are stems intercepted by a line cutting across that similarity level. 

 The agrument for doing so would be to maintain the objectivity of the data 

 analysis. 



It has been shown that the supposed objectivity of the data analysis is 

 not really there to maintain. Furthermore, any supposed objectivity gained at 

 this point is bought at a higher cost, for the ecologist must essentially ig- 

 nore anything he has learned in the course of the study, both in the field 

 and by reviewing the works of others. 



The approach used in this study was to use the dendrogram as the graphic 

 summarization of the relationships among samples, and presumably among com- 

 munities. Clusters of samples suggesting community types were examined as 

 groups, and in relation to adjacent clusters to determine such things as 

 whether the clusters were separated on the basis of total coverage despite the 

 fact that relative abundance of the dominance was essentially the same, or 

 was a reversal of dominance or the addition of a new dominant involved, and if 

 so, how could it be interpreted. The photographs of each sample were useful 

 in this comparison because sites as well as floristics were portrayed. 



Since field experience was very useful when classifying, one of the 

 principal data gatherers worked on the classification. After its completion, 

 the preliminary classification was discussed with local experts and compared 

 with published studies involving classification or recognition of types. As 

 a result, modifications were made in the classification, but the dendrogram 

 was always the frame of reference for any revision. Multivariate analysis did 

 not, however, supplant the ecologist, whose interpretation and understanding 

 of the phytocoenosis finally determined the classification. 



PRODUCTIVITY 



The productivity measurements used for reclamation purposes are rather 

 limited. Some definitions (Whittaker 1975) may be useful in explaining this. 



114 



