Changes in the soil weed seedbank were monitored over a single growing season in field experiments with drilled salad onion and carrot crops following different weeding treatments. Viable weed seed numbers were determined in soil samples taken at crop sowing and at harvest. When the crops were left weedy until harvest, seed numbers in soil increased by between two- and seventy-fold due to the influx of fresh seed. There was much less change in weed seed numbers in the soil when crops were kept weed-free until harvest or weeded once at 4, 5 or 6 weeks after 50% crop emergence. An estimate of the seedbank at harvest in the weeded treatments as a proportion of that of the unweeded control was derived using a log-linear model that included a covariate adjustment for seed numbers at sowing. There was little difference in the proportion remaining when crops were weeded once or kept weed-free throughout. The results were similar in carrot and salad onion crops grown in organic or conventional cropping systems. The findings highlight the risk of a rapid increase in the soil seedbank if weed control is neglected in a single crop. However, there was no indication that limiting weed control to one carefully-timed weeding would increase the potential weed population in subsequent crops.