Abstract
Herbage yield and a range of environmental variables were recorded over two growing seasons. The resulting data were subjected to a multiple regression analysis in order to identify the major environmental variable associated with herbage yield. Soil moisture, soil temperature and soil aeration appeared the most important environmental variables involved. Results suggested that the construction of useful 'predictive' models from cutting data using multiple regression is associated with a number of problems; the definition of the dependent variable and multicolinearity were major problems encountered.
