We sought to measure the residual dentin thickness (RDT) in the mesial roots of mandibular molars after instrumentation with Lightspeed and Gates-Glidden rotary instruments. Study design Thirty extracted, untreated human mesial roots of mandibular molars were separated from the distal roots and embedded in clear polyester resin. The roots were cut horizontally at 1, 4, and 7 mm short of the anatomic apex. The diameter of each mesiobuccal canal was measured by using a stereo measuring microscope at each level in the buccolingual and mesiodistal directions. The dentin thickness was measured in each level in the mesial, distal, buccal, and lingual directions. Sections were reassembled with a muffle. The canals were enlarged to the working length with Lightspeed rotary instruments, of which the average size used was a No. 50 file. The coronal third was flared with No. 2 Gates-Glidden reamers. Slices were separated again, and the RDT and canal diameters were measured. The minimal measured RDT after instrumentation at the 1-, 4-, and 7-mm levels was 0.70 +/- 0.28 mm, 1.04 +/- 0.18 mm, and 1.09 +/- 0.19 mm, respectively. The average diameter of the canals after instrumentation at the 1-, 4-, and 7-mm levels was 0.50 +/- 0.04 mm, 0.52 +/- 0.05 mm, and 0.74 +/- 0.08 mm, respectively. The canal diameter did not exceed one third of the root diameter at all levels. Root canal preparation of mandibular mesial roots with Lightspeed instruments to No. 50 in the apical third and Gates-Glidden reamers to No. 2 in the coronal third does not significantly decrease the RDT.