20-, 40-and 60-cm) and weeds (presence or absence) on soybean isoflavone concentrations and isoflavone yield. Total and individual isoflavone concentrations were determined by HPLC. Seed yield, and oil and crude protein (CP) concentrations were concurrently determined. Year, seeding date, and weeds affected total and individual isoflavone concentrations, while row spacing had no effect. Total isoflavone concentration was 84% greater in 2003 than 2004. Seeding in mid-June increased isoflavone concentration by 38%, compared with seeding in May. The presence of weeds increased total isoflavone concentration by 9%. Year, row spacing, and weeds significantly affected seed yields. Seed yields were greatest in 2004, at 20-or 40-cm row spacing, and in the absence of weeds. Seeding date affected CP and oil concentrations. Greater CP concentration was observed with earlier seeding, the reverse was observed for oil. Weeds also affected CP and oil concentrations: higher CP and oil concentrations were observed in weedy and weed-free plots, respectively. Total isoflavone yield was affected by all factors evaluated. Isoflavone yield was greater in 2003 than 2004, with mid-June rather than late May seeding, when seeded at row spacing of 20-or 40-than 60-cm, and without weeds. Finally, negative correlations were observed between isoflavone concentrations and CP concentration and seed yield. It thus seems that certain agronomic practices may need to be tailored specifically to isoflavone production if concentrations in soybean are to be maximized. The negative correlations observed between isoflavone concentrations and other important seed characteristics warrant further research. La concentration de PB est plus élevée quand on sème plus tôt, l'inverse étant vrai pour l'huile. Les mauvaises herbes influent aussi sur la concentration de PB et d'huile, la première étant plus élevée dans les parcelles avec adventices et la seconde dans celles sans mauvaises herbes. Le rendement total en isoflavones subit l'influence de tous les paramètres examinés. Le rendement en isoflavones était plus élevé en 2003 qu'en 2004, avec les semis à la mi-juin plutôt qu'à la fin de mai, à un écartement de 20 ou de 40 cm plutôt que de 60 et dans les parcelles sans mauvaises herbes. Enfin, la concentration d'isoflavones présente une correlation négative avec le rendement grainier et celle de PB. Il semble donc qu'on devrait adapter certaines pratiques agricoles pour optimiser la production d'isoflavones par le soja. Les corrélations négatives observées entre la concentration d'isoflavones et d'autres caractéristiques importantes de la graine justifient la tenue d'autres recherches.