Herbicide trials on <i>Campuloclinium macrocephalum</i> (Asteraceae) during adverse conditions reveal incompatibility with biocontrol and a narrow window of opportunity for chemical control

Abstract

Herbicide trials on pompom weed Campuloclinium macrocephalum (Less.) DC. (Asteraceae) were carried out in grasslands with xeric and  hydric soils, at sites in Gauteng, South Africa, from 2005 to 2008. The rust fungus Puccinia eupatorii Dietel, an established self- perpetuating biocontrol agent on C. macrocephalum, caused significant damage to shoots in late summer and autumn. Herbicide  efficacy was significantly better on foliage that did not already present symptoms and that was sprayed in early summer compared with  on infected plants treated in autumn, at both the xeric (F = 36.71, p < 0.001) and hydric (F = 3.59, p = 0.031) sites. This pathogen has reduced the time available for effective chemical weed control from six months to three months. However, none of the herbicide  treatments achieved commercially acceptable mortality rates of ≥80%, and after three annual applications none succeeded in extirpating  C. macrocephalum from the plots. Regression analyses of count data showed that metsulfuron-methyl at 45 g ha^–1 (USD 20.05 ha^-1),  picloram at 252 g ha^–1 (USD 25.47 ha^–1) and 2,4-D/dicamba/MCPA at 540/360/474 g ha^–1 (USD 45.41 ha^–1) required five annual  treatments in summer to eradicate C. macrocephalum, whereas 2,4-D amine at 1 440 g ha^–1 (USD 13.65 ha-1) and MCPA at 1 200 g ha^–1  (USD 14.46 ha–1) required six treatments. Hence, there is a trade-off between the use of slightly more effective herbicides against the  lower costs of other selective herbicides.