Newete, Solomon Wakshim2010-07-302010-07-302010-07-30http://hdl.handle.net/10539/8334The conservative method of host specificity testing dictates that a potential biological control agent which shows polyphagous behaviour in the laboratory will be rejected, even though in a natural situation it may be monophagous or nearly so. To distinguish one from the other the performance of eucalyptus weevil, (Gonipterus scutellatus) was tested on 14 Eucalyptus and one Syzygium species in the laboratory, and the field. The weevil revealed different levels of polyphagy, depending on how the host plants were presented; as cut leaves, bouquets or sleeved-branches; or in choice or no-choice combinations. However, the fundamental host range was broader than the realized host range. Eucalyptus smithii and E. urophylla were the most preferred hosts (contrary to the literature), while E. saligna and Syzygium myrtifolia were immune to feeding and oviposition. Nevertheless, adult feeding and oviposition was more selective in the field, and the larvae are less discriminating than the adults. Finally, the weevil is shown to have a narrow host range within two sections of the subgenus Eucalyptus, sufficiently restricted if it was ever to be considered as a biocontrol agent.enThesis