The authors previously found that measurements of precorneal tear film thickness in animals, using laser interferometry and confocal microscopy, were larger than earlier estimates. They proposed that this occurred because optical methods did not disturb film structure and measured full thickness, including mucus. In the present study, tear film thickness was measured in humans. Coherent light was reflected from eyes and formed interference patterns. Thickness was determined from spacing of fringes. Mean thickness in six subjects was 34-45-mu-m, more than four times larger than earlier values. Validity and accuracy of measurements by interferometry were examined in our earlier study using confocal microscopy. Here it provided supportive evidence. Tear film thickness was estimated from optical sections through the corneas of three subjects to be 41-46-mu-m. Mucus content of the film also was examined. Interferometry was used to measure thinning after application of a mucolytic agent. There was no change in thickness after 5 min expsoure to 0.1% (weight/volume) acetylcysteine. Solutions of 20% caused thinning to 11-mu-m, and thickness slowly increased to reach its original values over 40 min. Thus, the film seemed to be composed substantially of mucus, not aqueous fluid. The results also provided evidence that measurements by interferometry did not include underlying epithelium. Earlier nonoptical methods probably did not include the portion of the film that contained mucus. They seem to have greatly underestimated human tear film thickness.