Aspects of the etiology of subcutaneous spindle‐cell neoplasms in feral gizzard shad (Dorosoma cepedianum) collected from Lake of the Arbuckles, Oklahoma, were investigated. Field collections of adult and juvenile gizzard shad were conducted to determine the prevalence and seasonality of neoplasms and site specificity of neoplasm‐bearing fish. Overall neoplasm prevalence for adult gizzard shad was 22.1%; no juveniles exhibited neoplasms. Neoplasm occurrence did not appear seasonal. Neoplasm‐bearing gizzard shad were found at all three sampling sites within Lake of the Arbuckles with similar neoplasm prevalence rates (Guy Sandy Creek 22.1%, Rock Creek, 17.9%, and Buckhorn Creek 22.4%). No gizzard shad (juvenile or adult) collected from Lake Carl Blackwell, Oklahoma, the reference site, exhibited neoplasms. Water, sediment, and shad liver tissue from Lake of the Arbuckles and the reference site were analyzed for total recoverable metals (cadmium, chromium, copper, nickel, and lead, by graphite furnace atomic absorption). Chromium, copper, and nickel were found in the water samples at concentrations of >1 to 8.6 μg/L. Low concentrations (>1 to 13.6 μg/g wet weight) of all the metals were found in the sediment and liver tissue from both sites. Heavy metal contamination does not appear linked to neoplasm occurrence. To determine if a nonbacterial infectious agent was responsible for tumor formation, cell‐free neoplasm homogenates were injected into healthy rainbow trout. Neoplasms were not transmitted into rainbow trout. The etiology of these neoplasms remains unknown.