Objective. This study determined the ultimate tensile strength (UTS) of enamel (E), dentin (D) and dentin-enamel junction (DEJ) using the microtensile technique. It was hypothesized that the UTS of dental structures varies according to location and nature. Methods. Intact occlusal, enamel surfaces from extracted human third molars were etched with 37% phosphoric acid and bonded with a one-bottle adhesive system. The bonded occlusal surfaces received a resin composite build-up and teeth were serially, vertically sectioned into several 0.7 mm thick stabs. Each stab was then trimmed to a dumbbell-shaped specimen with irrigated diamond burs to reduce the cross-sectional area to approximately 0.5 mm(2) at E, D or DEJ. E was tested according to its prismatic orientation (parallel, EP; and transversally, ET) and D as function of depth (superficial, DS; middle, DM and deep, DD). Specimens were tested in tension in an Instron testing machine at 0.5 mm/min. Results were analyzed by one-way ANOVA and Duncan's Multiple Range test. Results. UTS mean values (N = 20) were, MPa (SD): DEJ, 46.9 (13.7)(b); EP, 42.1 (11.9)(b); ET, 11.5 (4.7)(d); DS, 61.6 (16.2)(a); DM, 48.7 (16.6)(b) and DD, 33.9 (7.9)(c). Enamel stressed transversally to its prismatic orientation was significantly weaker (p < 0.05). Dentin depth significantly affected its UTS (p < 0.05). DEJ presented UTS that was similar to EP and DM (p > 0.05). Significance. The UTS of dental structures varies according to its nature and location. (C) 2003 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.204322329