The aim of this study was to analyze whether there are surnames which appear more frequently among the ancestors of cancer cases in a small isolate, in comparison to the ancestral surnames of the healthy controls, using the classic case-control design. The chosen setting was the island of Lastovo, Croatia, located more than 100 kilometers from the nearest coastal region. The period of study was 1970-1995, during which a total of 76 cancer cases were recorded in a population of approximately 800. The comparison of surname frequencies was performed in current and in five ancestral generations. The leading hypothesis was that, if inbreeding and common ancestry contributed to the development of the disease, then those phenomena should be reflected in increasing frequency of some surnames among ancestors, identifying the 'hidden' consanguinity, or 'following' cancer-promoting genes on the Y-chromosome. The results imply that there are surnames representing a classic "risk" for cancer, but also those "protecting" from its development, which all underscores the importance of founder effect and genetic predisposition to the disease in a small, reproductively isolated population. All of the results become more evident and increasingly significant when analyzed in more distant ancestral generations.