1. Different wheat cultivar samples were grown in replicated trials on a single site in three harvest years. A total of 23 wheat samples were harvested. Nutritionally complete, meal-form diets that included each wheat sample at 650 g/kg were used to compare broiler growth performance and determine apparent metabolisable energy (AME). The relationships were examined between these variables of nutritive value for broilers and the chemical composition and tests of quality on the wheat samples. 2. The total starch contents of the wheat samples ranged from 594 to 732 g/kg dry matter (DM). The mean total non-starch polysaccharide (NSP) content was 100g/kg DM comprising 73.5 g/kg DM of insoluble material. Endosperm hardness ranged between 10 and 77 relative units and Beaver was the softest cultivar in two of the batches of samples. The AME of the wheat-based diets had a maximum range of 0.5 MJ/kg within each of the three batches of wheat that were tested. 3. Endosperm hardness of the wheat grains, within a harvest year, was positively correlated with broiler weight gain and the 1000-grain weight was negatively correlated with feed conversion efficiency (FCE). Starch content and gross energy were positively correlated with determined AME, and DM, 1000-grain weight and water holding capacity were negatively correlated with AME. 4. A step-wise regression technique indicated that endosperm hardness and ash content of the wheat were the explanatory variables that, within harvest years, significantly reduced the unexplained variation in broiler growth rate, feed intake and FCE. The contents of total starch, crude protein and ether extract were the explanatory variables that, within harvest years, significantly reduced the unexplained variation in AME.