Interpreting ophiolites in terms of their original tectonic settings of formation has evolved through the interplay of marine and on-land research. Before 1960, the common occurrence of 'ophiolitic rock types' had been recognised and linked to submarine processes and thereby to the geosyncline concept. The 1960s was marked by the recognition that the best-preserved ophiolites had the structure and rock types of oceanic crust and underlying mantle, and this culminated, in 1972, with the Penrose Conference definition of ophiolites which is still used today. In the 1970s, the assumption that ophiolites formed at mid-ocean ridges began to be seriously tested when geochemical fingerprinting started to indicate a range of tectonic settings. This, and the rapid expansion of marine geology and ocean drilling, emphasised that spreading ridges in marginal ocean basins were, in many cases, better analogues than mid-ocean ridges for ophiolite formation and led to the use of the term 'supra-subduction zone ophiolite' for ophiolites with subduction components in their geochemical signatures. Development of the subduction zone-ophiolite link was a major theme of the 1980s, supported by many new cruises and a second major phase of ocean drilling in the Western Pacific. This enabled the petrological and geochemical features of marginal basin lithosphere to be properly defined. The science of the 1990s extended the likely range of modern analogues of ophiolite complexes. MARGINS science highlighted the relationship between many ophiolites and the lithosphere formed by a combination of tectonic and magmatic processes at rifted continental margins. A combination of ocean drilling and on-land studies revealed that ridge subduction could give a distinctive ophiolite-like sequence in a fore-arc setting. Ongoing work on Western Pacific drill-core emphasised the potential importance of subduction initiation for producing ophiolites, especially those containing boninitic magma types, and demonstrated that many marginal basins are floored by thinned arc crust, which resembles some ophiolitic sequences. The RIDGE campaign of the 1990s was also instrumental in highlighting differences between ultra-slow, slow- and fast-spreading ridges and so enabling still more detailed refinement of original tectonic setting.