The field experiments conducted on the grey-brown podzolic soil in the four growing seasons (1998–2001) at Krzeslice Farm, central-western Poland comprised seven fertilization variants: 80NF + 80CAN; 80CAN + 80CAN; 80AN + 80AN; 80NF + 50CAN + 30CN; 80CAN + 50CAN +30CN; 80AN + 50AN + 30CN (where NF – nitrofos NPK; CAN – calcium-ammonium nitrate; AN – ammonium nitrate; CN – calcium nitrate) and control (without N) applied in split rates at the beginning of spring regrowth (80 kg N ha−1), stem elongation (80 or 50) and flower buds visible stages (30). The yielding effect of tested fertilization variants was significant in comparison with the control (2.24 t ha−1). The highest mean seed yield (3.64 t ha−1) was collected from 80AN + 80AN and 80CAN + 80CAN variants. Mean values of 4 years indicate that the second N rate division (80 + 50 + 30) decreased yield, although not significantly in comparison with these two N treatments. Plants grown on these treatments have developed different patterns of growth to yield the seeds. These patterns were characterized by very high crop growth rate during flowering (above 21 g m−2 day−1) and negative at maturation (down to −2.5 g m−2 day−1). Plants fertilized with ammonium nitrate (80AN + 80AN) reached maximum growth rate earlier (65 days), which lasted longer (20 days) than plants fertilized with calcium-ammonium nitrate (71 days lasting 17.5 days). Plants grown on the control treatment reached the highest crop growth rate within 79 days (14.8 g m−2 day−1), which lasted 15 days.