The morphology of the glenoid cavity is highly variable, and no consensus exists regarding how to classify the different forms. We examined 98 dry scapulae to identify a common morphological entity and to define reproducible bony references of the glenoid cavity. The glenoid cavities were photographed perpendicularly in a standardized fashion. The bony peripheral rim was studied on these two-dimensional images, defined by randomly chosen points in order to define one or more circles. This study showed that only the peripheral rim of the inferior quadrants of the articular surface was found to be located on a circle ( P=0.926) with a radius of 12.8 mm (SD 1.3 mm). Defining the center of this circle appeared to be more reliable (ICC 0.98) than determining the middle point of the longitudinal axis (0,0) between the most cranial and most caudal points, defined as Saller's line (ICC 0.89). The distance of the center of this projected circle to the middle point of Saller's line had a unimodal distribution, suggesting the existence of only one glenoid cavity morphotype. We then investigated the relationship between the center of the circle and the area of subchondral bone thickening under the bare spot, the so-called tubercle of Assaki. Ten phenolized cadaveric glenoid cavities were examined with computed tomography. A circle was projected on the first image showing the bony peripheral rim, and this circle was copied on the consecutive slices until the tubercle of Assaki came across. The center of the circle was located within the area of the tubercle of Assaki, in all but one specimen. To investigate the clinical implications of this finding, the cadaver specimens were used to compare the position of the center of the circle with the postulated center of implantation according to the literature, and to the reference guide for a commonly used total shoulder prosthesis. The center of the circle was consistently situated more distal than the postulated center of the guide (mean 5.5 mm, range 4-8 mm) and the middle point of the glenoid cavity (mean 2 mm, range 1-3 mm). These findings could offer a reproducible point of reference for the glenoid cavity in osseous anthropometry and a valuable reference in shoulder replacement surgery, and might help in the definition of osseous glenohumeral instability.