The aim of the present experiment was to test a novel model system, designed to obtain human specimens of regenerated and also newly regenerated jaw bone, for the study of the biological events under a variety of conditions. Following information and disclosure of possible risks associated with a minor oral surgical procedure, 9 systemically healthy subjects (5 men, 4 women, mean age 31.7 years) signed consent forms and volunteered to participate in this study. Hollow test cylinders with an outer diameter of 3.5 mm, an inner diameter of 2.5 mm, and 4 mm in height were used. They were manufactured from commercially pure titanium and exhibited a highly polished inner surface and a titanium plasma sprayed outer rough surface. A mucoperiosteal flap was raised in the retromolar area of the mandible corresponding to standard retrained third molar surgery. Following flap reflection a standardized hole was drilled through the cortical bone into the bone marrow using round burs. The congruent test cylinders were firmly placed into the prepared bony bed yielding primary stability. One-and-a-half to 2 mm of the test device were submerged below the level of the surrounding bone, while the remainder surpassed the level of the bone surface. The bone-facing end of the cylinder was left open, while the coronal soft tissue facing end was closed by an ePTFE-membrane. The flap was sutured to obtain primary wound closure. In order to prevent infection, penicillin was prescribed systemically and oral rinses of chlorhexidine were administered. After 2, 7, and 12 weeks one test device including the regenerated tissue was surgically harvested, while after 16, 24 and 36 weeks respectively, 2 devices were harvested and processed for soft or hard tissue histology or histochemistry. The two surgical procedures and the presence of the test cylinders during the time of healing were well tolerated by the volunteers. In all 9 subjects generated tissue could successfully be harvested. The tissue generated after 2 and 7 weeks presented with a cylindrical shape, whereas the specimens harvested at 12 weeks and thereafter resembled the form of an hourglass. Specimens of 12 weeks and less regeneration time were almost entirely comprised of soft tissue, while specimens with regeneration time of 4 months and more were composed of both soft and increasing amounts of mineralized tissue. It is concluded that the presented model system is suitable to study temporal dynamics and tissue physiology of bone regeneration in humans with minimal risk for complications or adverse effects to the volunteers.