This study compared the precision and accuracy of 4 tear sampling methods. In vivo, albino rabbits were treated with single bilateral eye drops ofofloxacin 0.3% solution, 3 hr after which tear samples were collected using capillary tubes (CT), surgical sponges (SS), or tear strips for 15 sec (15sTS) or 60 sec (60sTS). In vitro, CT, SS, and tear strips were spiked with known volumes of ofloxacin solution in order to assess the bioanalytical accuracy of each technique. Ofloxacin levels were quantified by HPLC in all samples. Results showed that tear volumes and ofloxacin masses sampled in vivo depended on sampling method. Tear volume followed the rank order 60sTS > 15sTS > SS > CT. The volume collected by 60sTS exceeded precorneal tear volume. Ofloxacin mass followed the order 60sTS approximately 15sTS > SS > CT. Tear concentrations (mean +/- SD; N = 12) were 3.28 +/- 3.76 microg/g for 60sTS, 10.3 +/- 10.0 microg/g for 15sTS, 9.75 +/- 8.04 microg/g for SS, and 5.83 +/- 3.35 microg/g for CT. In vitro, SS, 15sTS, and 60sTS yielded accuracies of 103-107% and coefficients of variation (CV) < 9%. CT was only 85% accurate with a CV of 31%, indicating incomplete extraction during analysis. We concluded from this study that: 1) rabbit tear sampling by SS was rapid, easy, accurate, precise, and easily analyzable; 2) sampling by CT or 15sTS was accurate, but may require aggressive extraction (for CT) or be affected by tear flow rate (for 15sTS); and 3) tear sampling by 60sTS underestimated actual tear concentrations.