SUMMARYEleven carrot cultivars were tested to compare their susceptibility to carrot fly attack at Wellesbourne and at Mepal, Cambridgeshire in 1969. Significant differences in the percentages of unattacked roots were observed but not all the cultivars behaved consistently in this respect at the two sites. Foliage height and the percentage unattacked roots were not significantly correlated for these cultivars.Four of the cultivars representing the different levels of susceptibility were retested in fen and in mineral soil in microplots at Wellesbourne in 1970. Soil type significantly affected plant growth but not the relative susceptibilities of the cultivars to carrot fly attack. Plant size was positively correlated with insect attack and accounted for most, but not all, of the differences in susceptibility between the cultivars. In 1971 also, cv. Royal Chantenay was significantly less attacked than Speed's Norfolk Giant from August to December even after allowing for the large differences between the two cultivars in plant size.The experiments demonstrated a consistent difference in the relative susceptibilities of cvs Royal Chantenay and Speed's Norfolk Giant. The difference was correlated with plant size but allowances for foliage and root size failed to account for all of the difference. The results suggested a more fundamental basis for part of the difference in the susceptibility of the cultivars to carrot fly attack.