In four field experiments, the effects of single nitrogen (N) applications at planting on yield and nitrogen uptake of potato (Solanum tuberosum L.) was compared with two or three split applications. The total amount of N applied was an experimental factor in three of the experiments. In two experiments, sequential observations were made during the growing season. Generally, splitting applications (up to 58 days after emergence) did not affect dry matter (DM) yield at maturity and tended to result in slightly lower DM concentration of tubers, whereas it slightly improved the utilization of nitrogen. Maximum haulm dry weight and N content were lower when less nitrogen was applied during the first 50 days after emergence (DAE). The crops absorbed little extra nitrogen after 60 DAE (except when three applications were given). Soil mineral N (0–60 cm) during the first month reflected the pattern of N application with values up to 27 g/m2 N. After 60 DAE, soil mineral N was always around 2–5 g/m2. The efficiency of N utilization, i.e. the ratio of the N content of the crop to total N available (initial soil mineral N+deposition+net mineralization) was 0·45 for unfertilized controls. The utilization of fertilizer N (i.e. the apparent N recovery) was generally somewhat improved by split applications, but declined with the total amount of N applied (range 0·48–0·72). N utilization and its complement, possible N loss, were similar for both experiments with sequential observations. Separate analysis of the movement of Br− indicated that some nitrate can be washed below 60 cm soil depth due to dispersion during rainfall. The current study showed that the time when N application can be adjusted to meet estimated requirements extends to (at least) 60 days after emergence. That period of time can be exploited to match the N application to the actual crop requirement as it changes during that period.