Blood samples were obtained through the Blood Transfusion Service in Ireland in order to obtain information on the prevalence of asymptomatic B. burgdorferi infections and in an attempt to identify the type of habitat that presents the most risk of infection. Areas in the country were rated as low, medium or high risk based on the availability of suitable tick habitat, access to the public and the distribution of deer, the latter parameter being related to both the occurrence of rodent reservoir hosts and woodland recreational areas. Approximately 100 plasma samples from each of four areas were analysed for IgG anti-Borrelia antibodies by indirect immuno-fluorescence with a titre of 1 :80 indicating a positive reaction in asymptomatic individuals. Prevalence figures of 15, 11, 8 and 5% were obtained for high, high/medium, medium/low and low risk areas respectively. No positive samples were detected in blood from an Icelandic population which is not exposed to I. ricinus bites. The overall subclinical prevalence (9.75%) is surprisingly high in view of the apparent rarity of clinical cases in Ireland, though under-diagnosis probably occurs. These results seem to indicate that farmland is less important than woodland as
Lyme borreliosis habitat. If this is so, it is probably due to the presence in woodland of Apodemus sylvaticus, a putative reservoir host, and also to the use of such areas for recreation at certain times of the year.