The ABO records of 392 blood donors in the Isle of Bute, Scotland, showed a significant excess of group A and decrease of group O compared with data from the adjacent mainland. A survey by questionnaire to explain these observations assessed the representative nature of the donor panel as a sample of the Bute population, with respect to sex, age, occupation and origin by birthplace and parental birthplace. The reported motivations first prompting the Bute panel to donate blood were analysed, especially those liable to introduce bias into the results. It was concluded that: the informative 318 donors were not an atypical sample; inclusion of Younger Kindred enhanced the group A frequency; inclusion of donors with knowledge of their blood groups before enrolment as donors introduced a counter-bias in favour of group O; the raised group A frequency of the Incomers to the island was probably attributable to admixture of an English population with a non-Bute Scottish population; the Brandanes could represent a small indigenous island population in which the incidence of group A has become high, but, as numbers were small, a chance observation could not be excluded.