Colour difference vector analysis provides useful and meaningful information in scoring the Farnsworth-Munsell (FM) 100-hue test. However, the FM 100-hue test is limited in its ability to diagnose type and severity of congenital colour vision defect. Type classification for all subjects is incorrect in 21% of cases, and for deuteranomals the misclassification rate is 38%. Visual inspection of the plots yields a similar misclassification rate and classification of plots with few errors (under 180) is generally less reliable. The FM 100-hue test has a limited ability to separate dichromats from anomalous trichromats. A test protocol based on joint D15 and FM 100-hue tests should pass 36% of anomalous trichromats and 26% of all colour defectives yet fail all dichromatic observers. We conclude that administering the FM 100-hue test is of less value than a combination of D15 panels (Standard D15 and L'Anthony's desaturated D15) in the clinical diagnosis of congenital colour defective observers. Our results for the FM 100-hue panel are similar to those reported previously by other investigators.