Purpose: Implant length, implant surface area, and crown-to-root (c/r) ratio and their relationship to crestal bone levels were analyzed in 2 groups of partially edentulous patients treated with sintered porous-surfaced dental implants. Materials and Methods: One hundred ninety-nine implants were used to restore 74 partially edentulous patients with fixed prostheses. Implants were categorized according to their length (¡°short¡± versus ¡°long¡±) and estimated surface area (¡°small¡± versus ¡°large¡±). ¡°Short¡± implants had lengths of 5 or 7 mm, while ¡°long¡± implants were either 9 or 12 mm in length. ¡°Small¡± implants had estimated surface areas of = 600 mm², while ¡°large¡± implants had estimated surface areas > 600 mm². Other data collected included c/r ratio (measured on articulated diagnostic casts), whether or not the implants were splinted, and standardized sequential radiographs. Results: The mean c/r ratio was 1.5 (SD = 0.4; range 0.8 to 3.0), with 78.9% of the implants having a c/r ratio between 1.1 and 2.0. Neither c/r ratio nor estimated implant surface area (small or large) affected steady-state crestal bone levels. However, implant length and whether the implants were splinted did appear to affect bone levels. Long implants had greater crestal bone loss (0.2 mm more) than short implants; splinted implants showed greater crestal bone loss (0.2 mm more) than nonsplinted ones. These differences were statistically significant. Discussion and Conclusions: Sintered porous-surfaced implants performed well in short lengths (7 mm or less) in this series of partially edentulous patients. The data suggested that long implants and/or splinting can result in greater crestal bone loss; longer implants and splinted implants appeared to favor greater crestal bone loss in this investigation. These conclusions are, of course, specific to the implants used and would not be relevant to other implant types.