In recurrent tonsillitis, the pathogenic bacteria are harbored in the tonsil core, and therefore cultures of superficial swab samples are not particularly accurate in identifying specific types of core bacteria. On the other hand, the results of fine-needle aspiration (FNA) cultures of core samples have been closely correlated with the findings of core cultures in excised tonsils, and both methods are far superior to surface swabbing. We conducted a prospective study to compare the accuracy of culture findings from tonsillar tissue obtained by surface swabbing, FNA sampling of the tonsil core in situ, and core sampling of the excised tonsil in children with recurrent tonsillitis. Our patient population was made up of 54 children-22 boys and 32 girls, aged 4 to 14 years (mean: 10.7)-who were undergoing elective tonsillectomy during a 1-year period. On the day of surgery, a surface swab, core FNA sample, and dissected core sample were obtained from each patient and sent for culture. Culture showed that the three methods were in agreement in 34 cases (63.0%). In 9 cases (16.7%) the surface swab culture grew different pathogens from those of the two core cultures, and in 3 other cases (5.6%) the surface swab culture was negative while the two core cultures were positive for the same pathogens. In all, the results of core FNA culture and dissected core culture were in agreement in 46 cases (85.2%); in only 4 cases (7.4%) did the core FNA culture fail to accurately identify the causative pathogens. Overall, the sensitivity and specificity of core FNA sampling were 100 and 50% respectively, compared with 82.9 and 30.8% for the superficial tonsillar swab. We conclude that routine culture of surface swab specimens in patients with chronic or recurrent tonsillitis is neither reliable nor valid. We recommend that core FNA sampling be considered the diagnostic method of choice since it can be done on an outpatient basis, it would reliably allow for culture-directed antibiotic therapy, and it could obviate the need for elective tonsillectomy in many cases. However, its feasibility as an office procedure in children remains to be determined.