ORDERING MULTIVARIATE OBSERVATIONS
Intercropping data consists of multivariate observations of two or more crop-yields at different treatments (situations). Here, the problem is ordering these treatments with respect to the yields. At present land equivalent ratio (LER) is used to order the treatments and finding out at which treatment the LER has maximum value, considering that treatment is best one, to those intercrops. But, two or more crop-yields are not necessarily additive but may be interactive, then applying of LER could be misleading. In the present paper, we propose an alternative approach which brings out natural ordering if any exists, among the treatments.