Objective: To compare in vitro the influence of 3 cementation techniques on leakage and fracture strength of alumina all-ceramic crowns after cyclic loading in an artificial chewing simulator. Method and Materials: Forty-eight extracted molars were mounted in a way that simulates natural tooth mobility. Crowns (Procera Alumina, Nobel Biocare) were fabricated and inserted with either conventional cementation with zinc phosphate cement without pretreatment (group ZOP); cementation with a universal adhesive resin cement without pretreatment (group HYB); or adhesive bonding with composite resin after pretreatment of the tooth (dentin bonding agent) and the crown (airborne-particle abrasion and a special ceramic priming agent containing adhesive monomers that bond to metal-oxide ceramics) (group ADH). All specimens were stored in artificial saliva and subjected to 1.2 million load cycles in a dual-axis chewing simulator (Willytec). Eight specimens per group were subjected to compressive load until failure, while the remaining 8 specimens were stained and sectioned for measuring of dye penetration. One-way ANOVA and Tukey HSD were used for statistical analyses (a = .05). Results: Cementation techniques were significantly different (P = .009) in regard to mean load at failure. Fracture strength was significantly greater for ADH (mean load at fracture, 2,782 ± 419 N) as compared to HYB (1,980 ± 270 N) or ZOP (1,788 ± 242 N). All groups differed significantly for leakage values (P .001), with ADH showing the lowest mean leakage (0.04 ± 0.07 mm), followed by HYB (0.96 ± 0.16 mm) and ZOP (2.44 ± 0.19 mm). Conclusion: Cementation technique affects fracture strength and leakage of all-ceramic molar crowns. Fracture strengths were well above natural chewing forces for all cementation methods. However, adhesive bonding significantly increased fracture strength and improved marginal seal of alumina crowns. Schlagwörter: all-ceramic, alumina crown, cementation, fracture strength, marginal leakage