J. Alan A. RENWICK, Nicole TARNOWSKY
Boyce Thompson Institute, Tower Road, Ithaca, NY 14853, USA
Avoidance of some crucifers by Pieris rapae (a crucifer specialist) has been explained by the presence of oviposition deterrents in the unacceptable plants. In addition, feeding deterrents have been found to offer a second line of defense in these plants. Erysimum cheiranthoides contains strophanthidin-based cardenolides that act as oviposition deterrents to P. rapae and digitoxigenin-based cardenolides that deter feeding by larvae of the same insect. Similarly, two cucurbitacin glycosides have been identified as oviposition deterrents from Iberis amara, and one of these also acts as a feeding deterrent. It is not known whether these deterrents function to warn the insects of toxicity or whether they are toxic themselves. This is a question that has been difficult to answer for any insect, since the deterrents effectively prevent ingestion of potentially toxic constituents. However, recent studies have shown that dietary experience can affect the sensitivity of P. rapae larvae to deterrents, thus affecting their ability to discriminate between foods of differing quality. When larvae are reared on wheat germ diet, their development of sensitivity to deterrents is suppressed. It is now possible, therefore, to precondition larvae so that they will feed on plants that they would normally avoid.
Larvae of P. rapae were reared on wheat germ diet and transferred to Erysimum cheiranthoides or to Iberis amara plants at different stages in their development. On E. cheiranthoides, individual iaNae of ali stadia died within a few days. On I. amara survival was longer, but eventually most larvae died. Some larvae never fed on the plants, and when these were transferred to cabbage after a period of starvation, they recovered and continued to develop normally.
Toxicity of E. cheiranthoides constituents was tested by applying extracts to small cabbage leaves and allowing diet-reared larvae to feed on the treated leaves. Experiments with the different larval stages indicated that 5th instars were most suitable for bioassays, since they consumed more and died sooner than the earlier instars. Using this assay system, aqueous extracts of E. cheiranthoides were fractionated by solvent partitioning, and the activity was detected in a n-butanol fraction. Subsequent flash chromatography on a C18 silica column resulted in the isolation of an active fraction containing the cardenolides that were previously identified as oviposition and feeding deterrents. However, some prominent cardenolides appeared to be inactive as toxins and a few relatively minor members of the group may account for most of the toxicity.
The results show that E. cheiranthoides is protected from herbivory by P. rapae in 3 ways. Cardenolides are responsible for deterring oviposition by the adults as well as feeding by the larvae. However, if feeding should occur as a result of previous dietary experience, ingestion of the same or additional cardenolides proves to be lethal. The behavioral response of P. rapae to deterrents in this plant clearly provides an adaptive advantage to the species by avoiding the ingestion of toxins. Similarly, the deterrents present in I. amara seem to serve as a warning to P. rapae of the toxic nature of this plant. The results seem to support the idea that taste is a good indicator of quality. However, it is known that mistakes often occur, and other examples will be needed to show a more definite link between deterrency and toxicity.