Four experiments were carried out over a 2-year period to investigate the effect of plant density and spatial arrangement and the time of lifting on yield and root size of carrots. Plant arrangement was varied by growing the plants in 1-, 2-, 3-, 4-, 5- and 10-row systems in beds with rows 12–5 and 37–5 cm apart. With each row arrangement crops were grown at target densities of 108, 323 and 537 plants/ma. There were three times of harvest. Total root yield was not significantly affected by plant arrangement or, in three out of the four experiments, by plant density but yields progressively increased with later harvests at all density levels. Yields of canning-size roots (20–30 mm diameter) were influenced by plant density and time of harvest and there were highly significant interactions between these variables on canning root yields. Highest absolute yields were obtained from the latest harvests from the medium and high plant density treatments; with the lowest density treatments the highest yields of canners were obtained from the earliest harvests. Mean root weight was significantly affected by plant density and harvest time. Row system effects on mean root weight were shown to be largely accounted for by differences in percentage seedling emergence and hence the plant density obtained under the different row system treatments and a possible reason for this effect is discussed. The relationship between mean root weight and plant density was not affected by the pattern of plant arrangement in these experiments. Total root yields were shown to be related to cumulative potential evapotranspiration from sowing rather than to chronological time. It is concluded that with the range of row systems used in these experiments, row spacing had little effect on root yield but confirmed previous work that plant density and time of harvest were the most important variables controlling mean root size and hence the yield of roots of any particular size grade.