APPENDIX C. CONSIDERATIONS AFFECTING STUDY PROCEDURES 



PERSPECTIVES ON CLASSIFICATION 



The Controversy 



Man has always classified not only artifacts, but natural phonemena , and 

 the classification of vegetation comes naturally to niany--so nat'jrally, in 

 fact, that many classifications have been proposed without any quanitative 

 data collection or analysis, and without stating the basis of the classifi- 

 cation. This tendency has been particularly strong in the United States, 

 where the broadly defined dominance types of Clements (1915, 1936) have been 

 influential . 



Ramensky (1926) and Gleason (1926) recognized the autonomous properties 

 of species and the role of chance in species composition. They refuted the 

 organismic theory of communities and took a more individualistic view. At 

 the time, their thoughts had little impact in the United States. These issues 

 resurfaced in the 1950's and 1960's. 



Mcintosh (1967) notes a fundamental question of classification asking 

 "how many intermediates must be interspersed between "types" before they are 

 lost as types and become continuous with one another". Are the classes re- 

 sulting from the classification of vegetation mere artifacts, the imposition 

 of a man-made order where there is none, or can classification be valid as well 

 as useful? Mcintosh (1967) gives a good introduction to the controversy. 



These questions resulted in some mathematical techniques for comparing 

 plant communities. Bray and Curtis (1957) started a movement toward ordin- 

 ation which has been taken up and refined by Beals (1960), Swan and Dix (1966), 

 Newsome and Dix (1968) and Swan, Dix and Wehrahn (1969), to name a few. 



Whittaker, (1956, 1967) and Whittaker and Niering (1965) favor gradient 

 analysis as an alternative to classification. 



Sokol and Sneath's textbook on numerical taxonomy (1963) started a move- 

 ment toward cluster analysis, which has been applied by Thelenius (1972) and 

 West (1966). 



Out of the controversy and the application of various techniques has 

 come a better understanding of plant communities and improved methods of com- 

 paring communities through multivariate analysis. The results of ordination, 

 gradient analysis and cluster analysis have major significance for under- 

 standing the practical problems of classification. 



In addition to questions concerning the validity and appropriateness of 

 classification, there is also the question of the need and usefulness of class- 

 ification. For example, Gilmour (1951) maintains that classification is the 

 prerequisite for all thought, Webb (1954) asserts that some system of typi- 



IC9 



