Root collar excavation for control of Armillaria root disease of grapevine was investigated in two California vineyards (vineyard K1 and vineyard N1) from 2002 to 2004. The hypothesis tested was that root collar excavation, when timed in early stages of root collar infection, may cause mycelial fans of the pathogen Armillaria mellea to recede from the root collar before severe disease results from vascular tissue decay. In vineyard N1, excavation significantly increased yield and cluster weight of symptomatic grapevines; symptomatic-excavated grapevines had the same high mean cluster weight as healthy grapevines, and there were no significant effects of excavation on yield or pruning weight of healthy grapevines. In vineyard K1, where excavated root collars frequently refilled with soil, excavation had no significant effects on yield or pruning weight of symptomatic grapevines, and significantly reduced pruning weight and shoot weight of healthy grapevines. Reexamination in March 2004 revealed that mycelial fans had receded from root collars of symptomatic-excavated grapevines, but remained on root collars of symptomatic-nonexcavated grapevines. Root collar excavation appears to be a promising cultural approach for control of Armillaria root disease, as long as excavated root collars are kept clear of soil.