AbstractThe leaf trichome densities of upper and lower surfaces of wild, Mediterranean Brassica spp. were determined using light and electron microscopic techniques. Two groups of plant species, one with a density of 2171–5312 trichomes per cm2 and the other with 0–30 trichomes per cm2, were tested in the laboratory for resistance against flea beetle feeding. The extent of feeding was determined in choice tests using leaf discs or detached whole leaves and in no-choice tests using clip-on cages on intact plants. Of the seven species tested, only two, B. villosa Biv. and B. villosa Biv. subsp. drepanensis which had a trichome density of > 2172 per cm2, were found to be highly resistant to flea beetle feeding. All other species had a trichome density of < 30 per cm2 and all suffered significant damage from flea beetle feeding. Behavioural observations indicated that a high density of trichomes on undamaged B. villosa leaves acted as a physical barrier to flea beetle feeding by preventing the flea beetles from firmly settling on the leaf surface to initiate feeding. Flea beetles readily initiated feeding on areas of B. villosa leaves where trichomes were inadvertently damaged suggesting that this plant did not contain significant concentrations of feeding deterrents. Electroantennogram (EAG) studies also indicated that B. villosa leaves are as stimulating to flea beetles as other Brassica species tested. Behavioural observations, feeding tests, and EAG studies indicated that trichomes, not other factors such as chemical repellents or feeding deterrents, are responsible for the resistance in B. villosa plants to flea beetle feeding.