OVIPOSITION AND CHEMOSENSORY STIMULATION OF THE ROOT FLIES DELIA RADICUMAND D-FLORALIS IN RESPONSE TO PLANTS AND LEAF SURFACE EXTRACTS FROMRESISTANT AND SUSCEPTIBLE BRASSICA GENOTYPES

In Brassica crops differences in susceptibility to root fly attack can be largely attributed to antixenotic resistance. Plants of four genot ypes (two swedes and two kales) with widely differing resistance in fi eld trials, were compared in laboratory choice assays for their suscep tibility to oviposition by the root flies Delia radicum (L.) and D. fl oralis (Fallen) (Diptera, Anthomyiidae). For both species the preferen ce among the genotypes corresponded to the susceptibility of the genot ypes in the field. The preference ranking in response to surrogate lea ves treated with methanolic surface extracts of the four genotypes was identical to the preference among potted plants, demonstrating that c hemical factors on the leaf surface mediate host preference for ovipos ition in these species. For both species of fly, glucosinolates are ma jor oviposition stimulants and for D. radicnm an additional, nonglucos inolate oviposition stimulant, presently called CIF, is known. We desc ribe a procedure for chromatographic separation of glucosinolates from CIF in leaf surface extracts. In oviposition-choice assays with D. ra dicum, the CIF-fractions of the two swede genotypes applied to surroga te leaves received a 1.8 and 4.6 times higher proportion of eggs than the respective glucosinolate-fractions, confirming the major importanc e of CIF as an oviposition stimulant. The genotype of swede that was p referred by both fly species in tests with plants and methanolic leaf surface extracts, also stimulated oviposition more in tests with the g lucosinolate-fractions or the CIF-fractions derived from the surface e xtracts, respectively. Thus, glucosinolates and CIF together account f or the observed preference among the genotypes and may also be respons ible for their susceptibility under field conditions. In the two kale genotypes the preference for plants or surface extracts differed from the preference among the corresponding glucosinolate- and CIF-fraction s, indicating that additional, as yet unknown chemical factors may als o be involved. For both groups of stimulants tarsal chemoreceptors all ow electrophysiological monitoring of glucosinolate- and CIF-activity in fractionated surface extracts. For D. radicum the chemosensory acti vity of both glucosinolate- and CIF-fractions corresponded to the resp ective behavioural activity in the oviposition preference tests, sugge sting that preference for oviposition among genotypes can be predicted from the electrophysiological activity of their fractions. The chemos ensory response of D. floralis, in particular to the CIF-fractions, wa s less pronounced than the response of D. radicum, indicating interspe cific differences in the perception of the major oviposition stimulant s. We discuss the potential application of electrophysiological techni ques in support of other screening methods used in breeding for root f ly resistance in Brassica crops.