Melanoma is often managed outside hospital settings, creating the potential for underreporting to cancer registries. To our knowledge, completeness of melanoma capture in cancer registries has not been assessed using external data sources since the 1980s. We evaluated the melanoma capture rate from 1993 to 2009 in a provincial cancer registry. We identified all melanoma diagnoses in pathology reports from a major community laboratory in Ontario, Canada. Pathologically confirmed diagnoses were linked to Ontario Cancer Registry (OCR) records using health insurance numbers. We calculated capture rates as the proportion of patients with melanoma confirmed by a pathology report, with a corresponding melanoma diagnosis in OCR. OCR captured 3,798 of 4,275 (88.8, 95 % confidence interval: 87.9, 89.8 %) invasive melanoma diagnoses over the 17-year period. Annual capture rates of 94 % or higher were found for over half the study period. Among all 29,133 melanoma diagnoses in OCR, 27.6 % were registered based on a pathology report alone, compared with 3.4 % for non-cutaneous malignancies. This suggests that comprehensive capture of melanoma cases by a provincial cancer registry is achievable using source data from community laboratories. There is a need for ongoing validation to ensure data remain accurate and complete to reliably inform clinical care, research, and policy.