Population and behavioural studies on Calycomyza eupatorivora spencer (Diptera : Agromyzidae), a biological control agent of Chromolaena odarata (L.) King and Robinson (Asteraceae) in South Africa.

Abstract

Chromolaena odorata (L.) King and Robinson (chromolaena, triffid weed) (Asteraceae: Eupatorieae) is one of the most problematic weeds in the subtropical northeastern parts of South Africa. Calycomyza eupatorivora Spencer (Diptera: Agromyzidae) was introduced as a biological control agent for the control of this weed. No study has yet been done to quantify field populations of C. eupatorivora since its establishment in 2003. The aim of this study was therefore to measure aspects of the field population and laboratory behaviour of C. eupatorivora on C. odorata. The first objective was to determine the percentage leaf area mined by larvae of C. eupatorivora on C. odorata plants exposed to three densities of mated flies, and also to determine the number of mines produced by these different densities, and their distribution on the plant. It also attempts to determine the relationship between chromolaena leaf quality and usage by C. eupatorivora. The maximum percentage of leaf area damaged was 37.5% for one of the trials involving five pairs of flies. Mean percentage leaf area damaged was slightly higher with five (28.5%) than ten pairs (22.0%) of adults and was lowest with one pair (6.5%), but these differences were not significant. In relation to the mean number of mines per plant, five and ten pairs of flies caused slightly more mines than one pair. The other significantly different parameter was number of leaves mined per plant, which was higher for five pairs. Within a plant, C. eupatorivora probably selects a subset of leaves with certain chemical and physical characteristics for oviposition since certain leaves were left unmined while others received multiple eggs. Percentage water content did not differ between mined and unmined leaves, but clear patterns were shown by acid detergent lignin which was higher in unmined leaves and nonstructural carbohydrates which were much higher in mined leaves. It is likely that leaf age plays a role in its suitability. The second objective was to quantify C. eupatorivora infestation levels, by counting and examining larval leaf mines, on C. odorata in the field at four times ('seasons' - September, December, March and July) over a 12-month period, and at three study sites that each included two habitats, viz. open and shady. At each of these six sampling sites, line transects were laid out and plants/branches sampled along them. Both plant/branch height and the number of leaves increased between September and March, and plants in the open habitats were taller and had more leaves than those in the shaded habitats. At the third site, the shady habitat supported taller plants with more leaves compared to the same habitat at the other sites. There was a steep increase in the number of C. eupatorivora mines from December to March. The mean number of mines, both total and in relation to leaves available, was highest in March, and was higher in the shaded habitats compared to the open habitats. The mean number of mines per damaged leaf was slightly higher in December compared to the other seasons, and was also higher in the open than the shaded habitats. Mean larval mortality was high (70%) in September but decreased to 32% in December, and increased again in late summer. The overall levels of mining by C. eupatorivora were low, with less than 5% of leaves sampled having mines. Taken together, the laboratory and field trials suggest that C. eupatorivora is restricted to a subset of the leaves of C. odorata for its development; that the field population is unable to make full use of the resource of young, palatable leaves that develop in early- to mid-summer because it only becomes large in late summer; and that the high mortality rate of young larvae negatively affects both the population of the fly and the level of damage to the plant. Given that these results were obtained in an area where the population of C. eupatorivora is relatively high, it is unlikely that the fly is having anything more than a negligible effect on C. odorata in South Africa at present.