The genetic models behind two approaches, graphical analysis due to Jinks and Hayman and combining ability analysis due to Griffing, employed in the analysis of data from diallel crosses are reviewed and their relative merits in plant breeding discussed. It is shown that the methods given by Griffing provide all the information that a breeder will need from a diallel cross. A few lapses that were noticed in published literature in the diallel cross experiments are brought out.
