This in vitro study compares the marginal adaptation of Class V restorations with margins located half in enamel and half in dentine, which were placed using different restorative techniques. Five operative procedures were evaluated both in saucer-shaped erosion lesions and in box-shaped cavities with bevels in enamel. The five procedures included a composite inlay technique using both the chemically and the light curing versions of a resin based composite cement, a bulk placement technique using a chemically curing composite resin, an incremental technique and an incremental technique combined with a built-up base, using a light curing composite resin. A combination of Gluma/Clearfil served as the dentinal adhesive. The micromorphology of the tooth/restoration interface was analysed before and after thermal cycling; the marginal seal was analysed after thermal cycling only. In the conventional cavities, the restorations showed less leakage, and micromorphologically a better, but statistically insignificant superior marginal adaptation. The inlay technique rendered the best marginal quality in both enamel and dentine before and after thermal cycling. Due to the unique curing characteristics of the chemically cured composite resin and cement resulting in a significantly reduced rigid contraction, the inlays cemented with the chemically curing cement and the restorations placed with the chemically curing composite resin were superior to their light cured counterparts. The built-up base yielding a reduction of the composite mass did not enhance marginal adaptation because of the partial replacement of the strong adhesion to dentine mediated by the Gluma/Clearfil combination by the weaker bond promoted by the etched glass ionomer cement.