1. The origins of agriculture, 10 000 years ago, led to profound changes in the biology of plants exploited as grain crops, through the process of domestication. This special case of evolution under cultivation led to domesticated cereals and pulses requiring humans for their dispersal, but the accompanying mechanisms causing higher productivity in these plants remain unknown. The classical view of crop domestication is narrow, focusing on reproductive and seed traits including the dispersal, dormancy and size of seeds, without considering whole-plant characteristics. However, the effects of initial domestication events can be inferred from consistent differences between traditional landraces and their wild progenitors. 2. We studied how domestication increased the yields of Fertile Crescent cereals and pulses using a greenhouse experiment to compare landraces with wild progenitors. We grew eight crops: barley, einkorn and emmer wheat, oat, rye, chickpea, lentil and pea. In each case, comparison of multiple landraces with their wild progenitors enabled us to quantify the effects of domestication rather than subsequent crop diversification. To reveal the mechanisms underpinning domestication-linked yield increases, we measured traits beyond those classically associated with domestication, including the rate and duration of growth, reproductive allocation, plant size and also seed mass and number. 3. Cereal and pulse crops had on average 50% higher yields than their wild progenitors, resulting from a 40% greater final plant size, 90% greater individual seed mass and 38% less chaff or pod material, although this varied between species. Cereal crops also had a higher seed number per spike compared with their wild ancestors. However, there were no differences in growth rate, total seed number, proportion of reproductive biomass or the duration of growth. 4. The domestication of Fertile Crescent crops resulted in larger seed size leading to a larger plant size, and also a reduction in chaff, with no decrease in seed number per individual, which proved a powerful package of traits for increasing yield. We propose that the important steps in the domestication process should be reconsidered, and the domestication syndrome broadened to include a wider range of traits.