Caribbean yellow-band syndrome is a highly prevalent coral disease whose transmission mechanisms are unknown. Affected corals often show multifocal lesions of yellow-colored tissue. We tested the hypothesis that a mechanical vector was responsible for these multifocal lesions. Four presumably non-resistant colonies were experimentally manipulated. Tissue and microbial assemblages were taken from diseased tissue and injected into healthy-looking tissue. Seawater injections were used as controls. The manipulations created a small wound, as would be generated by a coral predator. After 1 mo, all lesions healed and showed no signs of disease. We therefore reject the hypothesis that a mechanical vector, acting in a similar way as our manipulations, was responsible for the multifocal lesions.