During the 1990s, Roger Cowley had a strong interest in the crystal and magnetic structures of rare-earth superlattices as a means to understand the rich and exotic magnetic properties of the rare-earth metals. High-quality samples can be grown by molecular beam epitaxy on sapphire substrates by first depositing a thin epitaxial layer of niobium, then a layer of yttrium or lutetium as a seed. High-resolution x-ray scattering is an excellent probe to characterise the crystal quality and was used to study the structure of the niobium layer. However, relatively little attention was paid to the seed layer. This article summarises some of the x-ray experiments performed by the Cowley group to study the structure of epitaxial (110) niobium on (11 (2) over bar0) sapphire, and extends the work to report some results on the structure of thin (0 00 1) yttrium seed layers. The structure of the yttrium films is shown to have a strong dependence on the thickness of the niobium buffer, with the buffer needing to be thicker than a critical value of similar to 80 degrees A for the formation of misfit dislocations at the Nb/Al2O3 interface before highly coherent Y films can be grown. Yttrium films grown on Nb buffers thinner than similar to 500 degrees A show a similar two-peak line shape in q(perpendicular to) scans through their specular Bragg peaks to that seen in the specular Nb Bragg peaks, with a resolution-limited feature on a broader diffuse peak. The resolution-limited feature depends on the thickness of the yttrium film, becoming weaker and having a stronger decay with increasing q(parallel to) as the film thickness increases, while the width of the yttrium broad peak evolves as the square root of the width of the niobium Bragg peak. The data are discussed within the context of theories describing the scattering from films with misfit dislocations.