Coral diseases have been increasingly reported over the past few decades and are a major contributor to coral decline worldwide. The Caribbean, in particular, has been noted as a hotspot for coral disease, and the aptly named white syndromes have caused the decline of the dominant reef building corals throughout their range. White band disease (WBD) has been implicated in the dramatic loss of Acropora cervicornis and Acropora palmata since the 1970s, resulting in both species being listed as critically endangered on the International Union for Conservation of Nature Red list. The causal agent of WBD remains unknown, although recent studies based on challenge experiments with filtrate from infected hosts concluded that the disease is probably caused by bacteria. Here, we report an experiment using four different antibiotic treatments, targeting different members of the disease-associated microbial community. Two antibiotics, ampicillin and paromomycin, arrested the disease completely, and by comparing with community shifts brought about by treatments that did not arrest the disease, we have identified the likely candidate causal agent or agents of WBD. Our interpretation of the experimental treatments is that one or a combination of up to three specific bacterial types, detected consistently in diseased corals but not detectable in healthy corals, are likely causal agents of WBD. In addition, a histophagous ciliate (Philaster lucinda) identical to that found consistently in association with white syndrome in Indo-Pacific acroporas was also consistently detected in all WBD samples and absent in healthy coral. Treatment with metronidazole reduced it to below detection limits, but did not arrest the disease. However, the microscopic disease signs changed, suggesting a secondary role in disease causation for this ciliate. In future studies to identify a causal agent of WBD via tests of Henle-Koch's postulates, it will be vital to experimentally control for populations of the other potential pathogens identified in this study.