Cadaveric material is often used to guide and validate the interpretation of magnetic resonance imaging (MRI) studies. Direct correlation is achieved when the cadaver material is imaged and then directly sectioned in the plane of the image. Indirect correlation, where the cadaveric sections are compared with unrelated in-vivo images, is easier and more commonly used. Technical difficulties associated with the direct method include preservation of form and composition of the tissues, and the choice of the correct location and plane in which to section the cadaver. Using an MRI compatible ruler, designed to assist the direct correlation of MRI and cadaveric sections, we have examined 10 preserved (embalmed) cadaveric knees using different MRI sequences on several occasions. Despite these variations, subsequent sectioning of the cadaveric knees has shown good correlation with the MR images. Of 54 MR images compared with cadaveric sections, anatomical correlation was rated by independent observers as good or perfect in 47 (87%). This new, versatile and simple method can make better use of our preserved human cadaveric material and has potentially wide application; we are now developing it further to assess the technical capabilities of novel imaging sequences.