Examining the role of olfaction in dietary choice

1. Download : Download high-res image (136KB)
2. Download : Download full-size image

     

Correspondence Analysis of species by environmental variable matrices

Abstract. Correspondence Analysis of species by environmental variables matrices is a relatively simple method for handling matrices which include many species and variables. If columns represent species and rows represent the states in which the environmental variables have been divided, each element of the matrix shows the number of releves in which the simultaneous occurrence of that species and that environmental variable state have been recorded. In this paper the use of the method for presence-absence data is considered. The method is appropriate for studies whose main purpose is to look for broad correlations between groups of environmental variables and groups of species. It is particularly useful for handling heterogeneous data sets including rare species and species-poor releves, a situation in which Correspondence Analysis and Canonical Correspondence Analysis are frequently complicated by the production of axes explaining the relationship between rare species and species-poor releves. Examples are presented with artificial data as well as with real data collected in the Mexican portion of the Chihuahuan Desert (Bolsón de Mapimi). It is shown for the Mapimi data that the method is robust to the partition of variables into different numbers of states.