Canopy architecture is an easily measured visual trait in wheat that we have found to impact favourably on crop photosynthesis, biomass, grain yield, lodging, tillering and heat tolerance. The impact of canopy architecture on growth and yield is not surprising as there is a strong theoretical basis for canopy architecture influencing radiation interception and radiation use efficiency. Despite these advantages there is little evidence that canopy architecture is selected in breeding programs and it appears to be a trait neglected in breeding. We evaluated lines from a spring wheat MAGIC population that varies for canopy architecture as well as a small set of lines from Australian breeding programs including some released cultivars. Lines were selected only for extremes in canopy architecture and not for characteristics associated with grain yield. Experiments were grown over two years in Canberra, Australia and there were two sowing dates in each year. Trials were not irrigated. Canopy architecture of each plot was scored using a visual rating out of 10. Average grain yield in all experiments generally varied from 5 to 6 t ha-1. Grain yield from large bordered plots averaged over all trials was 10% higher in the erectophile lines than the planophile lines. This increase came from an increase in above-ground biomass and there was no difference in harvest index. Grain number per unit area was 26% higher but grain weight decreased by 9% in the erectophile lines compared with the planophile lines. Lodging was also significantly lower in lines with an erectophile canopy. There was little difference in flowering time or crop height when averaged over canopy type. Mapping was conducted on 1000 lines of a 4-way MAGIC population grown at an irrigated and a dryland site in NSW. Over 40 significant QTL were identified at the two sites demonstrating the complex genetics of canopy architecture. The QTL accounted for about 80% of the variation in canopy architecture at both sites. G x E was low as canopy architecture scores were highly repeatable across years (r2 = 0.85). This trait opens the way to improve wheat yields via crop biomass which in the past has been largely intractable.