The aim of this prospective study was to compare epidemiological and clinical data in patients with a positive Borrelia burgdorferi sensu lato culture and culture-negative erythema migrans skin lesions. Of the 546 adult patients with erythema migrans seen at our institution in 1997 in whom a skin biopsy was performed and the specimen cultured for the presence of B. burgdorferi sensu lato, 235 (43%) had a positive and 311 (57%) a negative skin culture. More women than men were present in both groups and women were also significantly older than men. Tick bites resulting in culture-positive erythema migrans predominated in May (p = 0.012), while in August and September tick bites with subsequent culture-negative skin lesions were more common (p = 0.018 and 0.011, respectively). Similarly, erythema migrans lesions noticed by our patients in May were significantly more often Borrelia culture positive than negative (p = 0.004), while lesions appearing in October were significantly more often culture negative (p = 0.004). In addition to these seasonal differences, the comparison of the large number of Borrelia skin culture-positive and -negative patients with erythema migrans also revealed differences in several clinical parameters including a larger diameter of skin lesions in the culture-positive group (p = 0.007 at presentation, and p = 0.039 at registration, respectively), a lesser number of multiple skin lesions (7/235 versus 27/311, p = 0.006), and a lower frequency of signs/symptoms (p = 0.039) associated with erythema migrans lesions in culture-positive than in culture-negative patients. We have no plausible explanation for the majority of these rather unexpected findings. Of the 59 patients who, prior to biopsy, had received brief courses of antibiotics known to be effective in the treatment of erythema migrans, 12 (20.3%) were culture positive. As anticipated, the ratio of culture positivity in pretreated patients was significantly lower (p < 0.001) than in those without antecedent antibiotic therapy.