Abstract:
The sweet potato butterfly (Acraea acerata Hew.) has become an important pest on sweet potato in Southern Ethiopia in the last two decades. This thesis deals with its ecology and economic importance in Southern Ethiopia.The insect is indigenous to this area where it feeds on both native wild lpomoea species and the introduced sweet potato (L batatas). However, its perfonnance was better on the new host, sweet potato. The small A. acerata populations on the scattered patches of the native lpomoea species evidently contribute little to the population dynamics of A. acerata in sweet potato fields. Several parasitoid species attack the immatures of A. acerata. Larvae are attacked by two parasitic wasps and a parasitic fly while a third parasitic wasp kills the pupae. Late instar larvae die from Beauveria bassiana infection. Mortalities inflicted by the natural enemies were usually low and they did not seem to act in a direct density dependent way.Enemies thus appear to have little effect during high population density. However, they may play a role in keeping the population at low density for a longer period. The insect breeds throughout the year with six discrete generations per year, which form generation cycles. An analysis of stage specific mortalities and recruitment rates from eighteen generations showed that no density dependent mortality operated in eggyoung larvae, young larvae-old larvae and old larvae-pupa stages. Density dependent mortality, however, operated in the interval pupa-adult and there was strong density dependence in recruitment rate. A probable reason for this finding is adult emigration from high-density populations. Butterfly population densities are temporally autocorrela ted with the partial autocorrelation function showing a positive one-generation lag and a negative two-generation lag which suggests that population regulation is a second order process with density dependence acting in the adult stages. A more detailed analysis, however, showed that larval density affects recruitment in the same generation,i.e. with almost no lag.A strong positive correlation was found between larval tent density, extent of defoliation and yield loss and the correlation was stronger for early than for late harvests.However, delayed harvesting did not reduce yield loss significantly. It is recommended to use insecticides at larval densities exceeding 4 tents/m2 during the first three months after planting. It is proposed that assessment of A. acerata damage should be based on young larvae density because they are easy to count, better correlated with yield loss and surveying at this development stage leaves enough time to decide on control measures.